def load_image(infile, n_channels): img_name, ext = os.path.splitext(os.path.basename(infile)) print("Input File: %s" % (img_name + ext)) img_L = util.imread_uint(infile, n_channels=n_channels) img_L = util.uint2single(img_L) img_L = util.single2tensor4(img_L) print('Brisque Score of input image : %f' % (brisq.get_score(infile))) return img_name, ext, img_L
def add_JPEG_noise(img): quality_factor = random.randint(30, 95) img = cv2.cvtColor(util.single2uint(img), cv2.COLOR_RGB2BGR) result, encimg = cv2.imencode( '.jpg', img, [int(cv2.IMWRITE_JPEG_QUALITY), quality_factor]) img = cv2.imdecode(encimg, 1) img = cv2.cvtColor(util.uint2single(img), cv2.COLOR_BGR2RGB) return img
def __getitem__(self, index: int) -> Dict[str, Union[str, torch.Tensor]]: # get H image img_path = self.img_paths[index] img_H = util.imread_uint(img_path, self.n_channels) H, W = img_H.shape[:2] if self.opt['phase'] == 'train': self.count += 1 # crop rnd_h = random.randint(0, max(0, H - self.patch_size)) rnd_w = random.randint(0, max(0, W - self.patch_size)) patch_H = img_H[rnd_h:rnd_h + self.patch_size, rnd_w:rnd_w + self.patch_size, :] # augmentation patch_H = util.augment_img(patch_H, mode=np.random.randint(0, 8)) # HWC to CHW, numpy(uint) to tensor img_H = util.uint2tensor3(patch_H) img_L: torch.Tensor = img_H.clone() # get noise level noise_level: torch.FloatTensor = torch.FloatTensor( [np.random.uniform(self.sigma[0], self.sigma[1])]) / 255.0 # add noise noise = torch.randn(img_L.size()).mul_(noise_level).float() img_L.add_(noise) else: img_H = util.uint2single(img_H) img_L = np.copy(img_H) # add noise np.random.seed(seed=0) img_L += np.random.normal(0, self.sigma / 255.0, img_L.shape) noise_level = torch.FloatTensor([self.sigma / 255.0]) img_H, img_L = util.single2tensor3(img_H), util.single2tensor3( img_L) return { 'y': img_L, 'y_gt': img_H, 'sigma': noise_level.unsqueeze(1).unsqueeze(1), 'path': img_path }
def __getitem__(self, index): # ------------------------------------ # get H image # ------------------------------------ H_path = self.paths_H[index] H_resolution = [int(s) for s in self.sizes_H[index].split('_')] img_H = _read_img_lmdb(self.h_env, H_path, H_resolution) img_H = util.uint2single(img_H) img_H = util.modcrop(img_H, self.sf) # ---------------------------------------------- # get down4 sharp image(img_h0) to be target in train for deblur # ---------------------------------------------- img_h0 = util.imresize_np(img_H, 1 / 4, True) # ------------------------------------ # get L image # ------------------------------------ L_path = self.paths_L[index] L_resolution = [int(s) for s in self.sizes_L[index].split('_')] img_L = _read_img_lmdb(self.l_env, L_path, L_resolution) img_L = util.uint2single(img_L) # ------------------------------------ # L/H pairs, HWC to CHW, numpy to tensor # ------------------------------------ img_deblur = [img_L,img_h0] img_deblur = util.generate_pyramid(*img_deblur,n_scales=3) img_deblur =util.np2tensor(*img_deblur) img_L = img_deblur[0] img_H = util.single2tensor3(img_H) img_L0 = img_L[0] return {'L0':img_L0, 'L': img_L, 'H': img_H, 'L_path': L_path, 'H_path': H_path, 'ls': img_L, 'hs': img_deblur[1]}
def __getitem__(self, index): # ------------------------------------ # get L image # ------------------------------------ L_path = self.paths_L[index] L_resolution = [int(s) for s in self.sizes_L[index].split('_')] img_L = _read_img_lmdb(self.l_env, L_path, L_resolution) img_L = util.uint2single(img_L) # ------------------------------------ # L/H pairs, HWC to CHW, numpy to tensor # ------------------------------------ img_deblur = [img_L] #,img_h0] img_deblur = util.generate_pyramid(*img_deblur, n_scales=3) img_deblur = util.np2tensor(*img_deblur) img_L = img_deblur[0] return {'L0': img_L[0], 'ls': img_L, 'L': img_L, 'L_path': L_path}
def __getitem__(self, index): H_path = 'toy.png' if self.opt['phase'] == 'train': patch_H = self.H_data[index] # -------------------------------- # augmentation - flip and/or rotate # -------------------------------- mode = np.random.randint(0, 8) patch_H = util.augment_img(patch_H, mode=mode) patch_H = util.uint2tensor3(patch_H) patch_L = patch_H.clone() # ------------------------------------ # add noise # ------------------------------------ noise = torch.randn(patch_L.size()).mul_(self.sigma / 255.0) patch_L.add_(noise) else: H_path = self.paths_H[index] img_H = util.imread_uint(H_path, self.n_channels) img_H = util.uint2single(img_H) img_L = np.copy(img_H) # ------------------------------------ # add noise # ------------------------------------ np.random.seed(seed=0) img_L += np.random.normal(0, self.sigma_test / 255.0, img_L.shape) patch_L, patch_H = util.single2tensor3(img_L), util.single2tensor3( img_H) L_path = H_path return {'L': patch_L, 'H': patch_H, 'L_path': L_path, 'H_path': H_path}
def main(): #0. global config #scale factor sf = 4 stage = 8 patch_size = [32, 32] patch_num = [3, 3] #1. local PSF #shape: gx,gy,kw,kw,3 all_PSFs = load_kernels('./data') #2. local model device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') model = net(n_iter=8, h_nc=64, in_nc=4, out_nc=3, nc=[64, 128, 256, 512], nb=2, sf=sf, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") #model.proj.load_state_dict(torch.load('./data/usrnet_pretrain.pth'),strict=True) model.train() for _, v in model.named_parameters(): v.requires_grad = True model = model.to(device) #positional lambda, mu for HQS, set as free trainable parameters here. ab_buffer = np.ones( (patch_num[0], patch_num[1], 2 * stage, 3), dtype=np.float32) * 0.1 ab = torch.tensor(ab_buffer, device=device, requires_grad=True) params = [] params += [{"params": [ab], "lr": 0.0005}] for key, value in model.named_parameters(): params += [{"params": [value], "lr": 0.0001}] optimizer = torch.optim.Adam(params, lr=0.0001, betas=(0.9, 0.999)) scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=1000, gamma=0.9) #3.load training data imgs_H = glob.glob('/home/xiu/databag/deblur/images/DIV2K_train/*.png', recursive=True) imgs_L = glob.glob('/home/xiu/databag/deblur/images/DIV2K_lr/*.png', recursive=True) imgs_H.sort() imgs_L.sort() global_iter = 0 N_maxiter = 200000 #def get_train_pairs() for i in range(N_maxiter): t0 = time.time() #draw random image. img_idx = np.random.randint(len(imgs_H)) img_H = cv2.imread(imgs_H[img_idx]) #img2 = imgs_L[img_idx] #img_L = cv2.imread(img2) #draw random patch from image #a. without img_L #draw random kernel PSF_grid = draw_random_kernel(all_PSFs, patch_num) patch_L, patch_H, patch_psf = draw_training_pair( img_H, PSF_grid, sf, patch_num, patch_size) #b. with img_L #patch_L, patch_H, patch_psf,px_start, py_start,block_expand = draw_training_pair(img_H, PSF_grid, sf, patch_num, patch_size, img_L) t_data = time.time() - t0 x = util.uint2single(patch_L) x = util.single2tensor4(x) x_gt = util.uint2single(patch_H) x_gt = util.single2tensor4(x_gt) k_local = [] for h_ in range(patch_num[1]): for w_ in range(patch_num[0]): k_local.append(util.single2tensor4(patch_psf[w_, h_])) k = torch.cat(k_local, dim=0) [x, x_gt, k] = [el.to(device) for el in [x, x_gt, k]] ab_patch = F.softplus(ab) ab_patch_v = [] for h_ in range(patch_num[1]): for w_ in range(patch_num[0]): ab_patch_v.append(ab_patch[w_:w_ + 1, h_]) ab_patch_v = torch.cat(ab_patch_v, dim=0) x_E = model.forward_patchwise_SR(x, k, ab_patch_v, patch_num, [patch_size[0], patch_size[1]], sf) loss = F.l1_loss(x_E, x_gt) optimizer.zero_grad() loss.backward() optimizer.step() scheduler.step() t_iter = time.time() - t0 - t_data print('[iter:{}] loss:{:.4f}, data_time:{:.2f}s, net_time:{:.2f}s'. format(global_iter + 1, loss.item(), t_data, t_iter)) patch_L = cv2.resize(patch_L, dsize=None, fx=sf, fy=sf, interpolation=cv2.INTER_NEAREST) #patch_L = patch_L[block_expand*sf:-block_expand*sf,block_expand*sf:-block_expand*sf] patch_E = util.tensor2uint((x_E)) show = np.hstack((patch_H, patch_L, patch_E)) cv2.imshow('H,L,E', show) key = cv2.waitKey(1) global_iter += 1 # for logging model weight. # if global_iter % 100 ==0: # torch.save(model.state_dict(),'./logs/uabcnet_{}.pth'.format(global_iter)) if key == ord('q'): break if key == ord('s'): torch.save(model.state_dict(), './logs/uabcnet.pth') torch.save(model.state_dict(), './logs/uabcnet.pth')
def main(): # ---------------------------------------- # Preparation # ---------------------------------------- model_name = 'usrnet' # 'usrgan' | 'usrnet' | 'usrgan_tiny' | 'usrnet_tiny' testset_name = 'set5' # test set, 'set5' | 'srbsd68' test_sf = [4] if 'gan' in model_name else [ 2, 3, 4 ] # scale factor, from {1,2,3,4} show_img = False # default: False save_L = True # save LR image save_E = True # save estimated image save_LEH = False # save zoomed LR, E and H images # ---------------------------------------- # load testing kernels # ---------------------------------------- # kernels = hdf5storage.loadmat(os.path.join('kernels', 'kernels.mat'))['kernels'] kernels = loadmat(os.path.join('kernels', 'kernels_12.mat'))['kernels'] n_channels = 1 if 'gray' in model_name else 3 # 3 for color image, 1 for grayscale image model_pool = 'model_zoo' # fixed testsets = 'testsets' # fixed results = 'results' # fixed noise_level_img = 0 # fixed: 0, noise level for LR image noise_level_model = noise_level_img # fixed, noise level of model, default 0 result_name = testset_name + '_' + model_name model_path = os.path.join(model_pool, model_name + '.pth') # ---------------------------------------- # L_path = H_path, E_path, logger # ---------------------------------------- L_path = os.path.join( testsets, testset_name) # L_path and H_path, fixed, for Low-quality images E_path = os.path.join(results, result_name) # E_path, fixed, for Estimated images util.mkdir(E_path) logger_name = result_name utils_logger.logger_info(logger_name, log_path=os.path.join(E_path, logger_name + '.log')) logger = logging.getLogger(logger_name) device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') # ---------------------------------------- # load model # ---------------------------------------- if 'tiny' in model_name: model = net(n_iter=6, h_nc=32, in_nc=4, out_nc=3, nc=[16, 32, 64, 64], nb=2, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") else: model = net(n_iter=8, h_nc=64, in_nc=4, out_nc=3, nc=[64, 128, 256, 512], nb=2, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") model.load_state_dict(torch.load(model_path), strict=True) model.eval() for key, v in model.named_parameters(): v.requires_grad = False number_parameters = sum(map(lambda x: x.numel(), model.parameters())) model = model.to(device) logger.info('Model path: {:s}'.format(model_path)) logger.info('Params number: {}'.format(number_parameters)) logger.info('Model_name:{}, image sigma:{}'.format(model_name, noise_level_img)) logger.info(L_path) L_paths = util.get_image_paths(L_path) # -------------------------------- # read images # -------------------------------- test_results_ave = OrderedDict() test_results_ave['psnr_sf_k'] = [] for sf in test_sf: for k_index in range(kernels.shape[1]): test_results = OrderedDict() test_results['psnr'] = [] kernel = kernels[0, k_index].astype(np.float64) ## other kernels # kernel = utils_deblur.blurkernel_synthesis(h=25) # motion kernel # kernel = utils_deblur.fspecial('gaussian', 25, 1.6) # Gaussian kernel # kernel = sr.shift_pixel(kernel, sf) # pixel shift; optional # kernel /= np.sum(kernel) util.surf(kernel) if show_img else None idx = 0 for img in L_paths: # -------------------------------- # (1) classical degradation, img_L # -------------------------------- idx += 1 img_name, ext = os.path.splitext(os.path.basename(img)) img_H = util.imread_uint( img, n_channels=n_channels) # HR image, int8 img_H = util.modcrop(img_H, np.lcm(sf, 8)) # modcrop # generate degraded LR image img_L = ndimage.filters.convolve(img_H, kernel[..., np.newaxis], mode='wrap') # blur img_L = sr.downsample_np( img_L, sf, center=False) # downsample, standard s-fold downsampler img_L = util.uint2single(img_L) # uint2single np.random.seed(seed=0) # for reproducibility img_L += np.random.normal(0, noise_level_img, img_L.shape) # add AWGN util.imshow(util.single2uint(img_L)) if show_img else None x = util.single2tensor4(img_L) k = util.single2tensor4(kernel[..., np.newaxis]) sigma = torch.tensor(noise_level_model).float().view( [1, 1, 1, 1]) [x, k, sigma] = [el.to(device) for el in [x, k, sigma]] # -------------------------------- # (2) inference # -------------------------------- x = model(x, k, sf, sigma) # -------------------------------- # (3) img_E # -------------------------------- img_E = util.tensor2uint(x) if save_E: util.imsave( img_E, os.path.join( E_path, img_name + '_x' + str(sf) + '_k' + str(k_index + 1) + '_' + model_name + '.png')) # -------------------------------- # (4) img_LEH # -------------------------------- img_L = util.single2uint(img_L) if save_LEH: k_v = kernel / np.max(kernel) * 1.2 k_v = util.single2uint( np.tile(k_v[..., np.newaxis], [1, 1, 3])) k_v = cv2.resize(k_v, (3 * k_v.shape[1], 3 * k_v.shape[0]), interpolation=cv2.INTER_NEAREST) img_I = cv2.resize( img_L, (sf * img_L.shape[1], sf * img_L.shape[0]), interpolation=cv2.INTER_NEAREST) img_I[:k_v.shape[0], -k_v.shape[1]:, :] = k_v img_I[:img_L.shape[0], :img_L.shape[1], :] = img_L util.imshow(np.concatenate([img_I, img_E, img_H], axis=1), title='LR / Recovered / Ground-truth' ) if show_img else None util.imsave( np.concatenate([img_I, img_E, img_H], axis=1), os.path.join( E_path, img_name + '_x' + str(sf) + '_k' + str(k_index + 1) + '_LEH.png')) if save_L: util.imsave( img_L, os.path.join( E_path, img_name + '_x' + str(sf) + '_k' + str(k_index + 1) + '_LR.png')) psnr = util.calculate_psnr( img_E, img_H, border=sf**2) # change with your own border test_results['psnr'].append(psnr) logger.info( '{:->4d}--> {:>10s} -- x{:>2d} --k{:>2d} PSNR: {:.2f}dB'. format(idx, img_name + ext, sf, k_index, psnr)) ave_psnr_k = sum(test_results['psnr']) / len(test_results['psnr']) logger.info( '------> Average PSNR(RGB) of ({}) scale factor: ({}), kernel: ({}) sigma: ({}): {:.2f} dB' .format(testset_name, sf, k_index + 1, noise_level_model, ave_psnr_k)) test_results_ave['psnr_sf_k'].append(ave_psnr_k) logger.info(test_results_ave['psnr_sf_k'])
def main(): #0. global config #scale factor sf = 4 stage = 5 patch_size = [32, 32] patch_num = [2, 2] #1. local PSF #shape: gx,gy,kw,kw,3 all_PSFs = load_kernels('./data') #2. local model device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') model = net(n_iter=5, h_nc=64, in_nc=4, out_nc=3, nc=[64, 128, 256, 512], nb=2, sf=sf, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") #loaded_state_dict= torch.load('./data/uabcnet_final.pth') loaded_state_dict = torch.load('./data/uabcnet_finetune.pth') model.load_state_dict(loaded_state_dict, strict=True) model.eval() for _, v in model.named_parameters(): v.requires_grad = False model = model.to(device) #positional lambda, mu for HQS, set as free trainable parameters here. ab_buffer = np.loadtxt('./data/ab_finetune.txt').reshape( (patch_num[0], patch_num[1], 2 * stage, 3)).astype(np.float32) #ab[2x2,2*stage,3] #ab_buffer = np.ones((patch_num[0],patch_num[1],2*stage,3),dtype=np.float32)*0.1 ab = torch.tensor(ab_buffer, device=device, requires_grad=False) ab = F.softplus(ab) #3.load training data imgs_H = glob.glob('./DIV2K_train/*.png', recursive=True) imgs_H.sort() global_iter = 0 N_maxiter = 1000 PSF_grid = using_AC254_lens(all_PSFs, patch_num) all_PSNR = [] out_folder = 'finetune' for i in range(N_maxiter): #draw random image. img_idx = np.random.randint(len(imgs_H)) img_H = cv2.imread(imgs_H[img_idx]) img_H = np.pad(img_H, [(12, 12), (12, 12), (0, 0)]) croppatch = imgpatch(img_H, 280, 280, 24) patches = croppatch.crop(img_H, 1) patch_E_list = [] patch_L_list = [] # patch_L,patch_H,patch_psf = draw_training_pair(img_H,PSF_grid,sf,patch_num,patch_size) for piece in range(len(patches)): patch_L, patch_H, patch_psf = draw_testing_pair( patches[piece], PSF_grid, sf, patch_num, patch_size) x = util.uint2single(patch_L) x = util.single2tensor4(x) x_gt = util.uint2single(patch_H) x_gt = util.single2tensor4(x_gt) k_local = [] for h_ in range(patch_num[1]): for w_ in range(patch_num[0]): k_local.append(util.single2tensor4(patch_psf[w_, h_])) k = torch.cat(k_local, dim=0) [x, x_gt, k] = [el.to(device) for el in [x, x_gt, k]] ab_patch = F.softplus(ab) ab_patch_v = [] for h_ in range(patch_num[1]): for w_ in range(patch_num[0]): ab_patch_v.append(ab_patch[w_:w_ + 1, h_]) ab_patch_v = torch.cat(ab_patch_v, dim=0) x_E = model.forward_patchwise_SR(x, k, ab_patch_v, patch_num, [patch_size[0], patch_size[1]], sf) patch_L = cv2.resize(patch_L, dsize=None, fx=sf, fy=sf, interpolation=cv2.INTER_NEAREST) patch_E = util.tensor2uint((x_E)) patch_E_list.append(patch_E[np.newaxis, :]) patch_L_list.append((patch_L[None, ...])) print(piece) img_E = croppatch.merge(patch_E_list) # croppatch_E = imgpatch(np.zeros_like(img_H), 256, 256, 0) # croppatch_E.crop(np.zeros_like(img_H),1) # img_E = croppatch_E.merge(patch_E_list) psnr, ssim = cal_psnrssim(img_E, img_H, 255) print(psnr) print(ssim) cv2.imwrite( os.path.join('./result', out_folder, 'resultE-{:04d}.png'.format(i + 1)), img_E) cv2.imwrite( os.path.join('./result', out_folder, 'resultH-{:04d}.png'.format(i + 1)), img_H) all_PSNR.append(psnr) #show = np.hstack((patch_H,patch_L,patch_E)) np.savetxt(os.path.join('./result', out_folder, 'psnr.txt'), all_PSNR)
def __getitem__(self, index): # ------------------------------------- # get H image # ------------------------------------- H_path = self.paths_H[index] img_H = util.imread_uint(H_path, self.n_channels) L_path = H_path if self.opt['phase'] == 'train': """ # -------------------------------- # get L/H/M patch pairs # -------------------------------- """ H, W = img_H.shape[:2] # --------------------------------- # randomly crop the patch # --------------------------------- rnd_h = random.randint(0, max(0, H - self.patch_size)) rnd_w = random.randint(0, max(0, W - self.patch_size)) patch_H = img_H[rnd_h:rnd_h + self.patch_size, rnd_w:rnd_w + self.patch_size, :] # --------------------------------- # augmentation - flip, rotate # --------------------------------- mode = np.random.randint(0, 8) patch_H = util.augment_img(patch_H, mode=mode) # --------------------------------- # HWC to CHW, numpy(uint) to tensor # --------------------------------- img_H = util.uint2tensor3(patch_H) img_L = img_H.clone() # --------------------------------- # get noise level # --------------------------------- # noise_level = torch.FloatTensor([np.random.randint(self.sigma_min, self.sigma_max)])/255.0 noise_level = torch.FloatTensor( [np.random.uniform(self.sigma_min, self.sigma_max)]) / 255.0 # --------------------------------- # add noise # --------------------------------- noise = torch.randn(img_L.size()).mul_(noise_level).float() img_L.add_(noise) else: """ # -------------------------------- # get L/H/sigma image pairs # -------------------------------- """ img_H = util.uint2single(img_H) img_L = np.copy(img_H) np.random.seed(seed=0) img_L += np.random.normal(0, self.sigma_test / 255.0, img_L.shape) noise_level = torch.FloatTensor([self.sigma_test / 255.0]) # --------------------------------- # L/H image pairs # --------------------------------- img_H, img_L = util.single2tensor3(img_H), util.single2tensor3( img_L) noise_level = noise_level.unsqueeze(1).unsqueeze(1) return { 'L': img_L, 'H': img_H, 'C': noise_level, 'L_path': L_path, 'H_path': H_path }
def main(): # ---------------------------------------- # Preparation # ---------------------------------------- noise_level_img = 30 # noise level for noisy image noise_level_model = noise_level_img # noise level for model model_name = 'ffdnet_color' # 'ffdnet_gray' | 'ffdnet_color' | 'ffdnet_color_clip' | 'ffdnet_gray_clip' testset_name = 'CBSD68' # test set, 'bsd68' | 'cbsd68' | 'set12' need_degradation = True # default: True show_img = False # default: False task_current = 'dn' # 'dn' for denoising | 'sr' for super-resolution sf = 1 # unused for denoising if 'color' in model_name: n_channels = 3 # setting for color image nc = 96 # setting for color image nb = 12 # setting for color image else: n_channels = 1 # setting for grayscale image nc = 64 # setting for grayscale image nb = 15 # setting for grayscale image if 'clip' in model_name: use_clip = True # clip the intensities into range of [0, 1] else: use_clip = False model_pool = 'model_zoo' # fixed testsets = 'testsets' # fixed results = 'results' # fixed result_name = testset_name + '_' + model_name border = sf if task_current == 'sr' else 0 # shave boader to calculate PSNR and SSIM model_path = os.path.join(model_pool, model_name + '.pth') # ---------------------------------------- # L_path, E_path, H_path # ---------------------------------------- L_path = os.path.join(testsets, testset_name) # L_path, for Low-quality images H_path = L_path # H_path, for High-quality images E_path = os.path.join(results, result_name) # E_path, for Estimated images util.mkdir(E_path) if H_path == L_path: need_degradation = True logger_name = result_name utils_logger.logger_info(logger_name, log_path=os.path.join(E_path, logger_name + '.log')) logger = logging.getLogger(logger_name) need_H = True if H_path is not None else False device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') # ---------------------------------------- # load model # ---------------------------------------- from models.network_ffdnet import FFDNet as net model = net(in_nc=n_channels, out_nc=n_channels, nc=nc, nb=nb, act_mode='R') model.load_state_dict(torch.load(model_path), strict=True) model.eval() for k, v in model.named_parameters(): v.requires_grad = False model = model.to(device) logger.info('Model path: {:s}'.format(model_path)) test_results = OrderedDict() test_results['psnr'] = [] test_results['ssim'] = [] logger.info('model_name:{}, model sigma:{}, image sigma:{}'.format( model_name, noise_level_img, noise_level_model)) logger.info(L_path) L_paths = util.get_image_paths(L_path) H_paths = util.get_image_paths(H_path) if need_H else None for idx, img in enumerate(L_paths): # ------------------------------------ # (1) img_L # ------------------------------------ img_name, ext = os.path.splitext(os.path.basename(img)) # logger.info('{:->4d}--> {:>10s}'.format(idx+1, img_name+ext)) img_L = util.imread_uint(img, n_channels=n_channels) img_L = util.uint2single(img_L) if need_degradation: # degradation process np.random.seed(seed=0) # for reproducibility img_L += np.random.normal(0, noise_level_img / 255., img_L.shape) if use_clip: img_L = util.uint2single(util.single2uint(img_L)) util.imshow(util.single2uint(img_L), title='Noisy image with noise level {}'.format( noise_level_img)) if show_img else None img_L = util.single2tensor4(img_L) img_L = img_L.to(device) sigma = torch.full((1, 1, 1, 1), noise_level_model / 255.).type_as(img_L) # ------------------------------------ # (2) img_E # ------------------------------------ img_E = model(img_L, sigma) img_E = util.tensor2uint(img_E) if need_H: # -------------------------------- # (3) img_H # -------------------------------- img_H = util.imread_uint(H_paths[idx], n_channels=n_channels) img_H = img_H.squeeze() # -------------------------------- # PSNR and SSIM # -------------------------------- psnr = util.calculate_psnr(img_E, img_H, border=border) ssim = util.calculate_ssim(img_E, img_H, border=border) test_results['psnr'].append(psnr) test_results['ssim'].append(ssim) logger.info('{:s} - PSNR: {:.2f} dB; SSIM: {:.4f}.'.format( img_name + ext, psnr, ssim)) util.imshow(np.concatenate([img_E, img_H], axis=1), title='Recovered / Ground-truth') if show_img else None # ------------------------------------ # save results # ------------------------------------ util.imsave(img_E, os.path.join(E_path, img_name + ext)) if need_H: ave_psnr = sum(test_results['psnr']) / len(test_results['psnr']) ave_ssim = sum(test_results['ssim']) / len(test_results['ssim']) logger.info( 'Average PSNR/SSIM(RGB) - {} - PSNR: {:.2f} dB; SSIM: {:.4f}'. format(result_name, ave_psnr, ave_ssim))
def main(): """ # ---------------------------------------------------------------------------------- # In real applications, you should set proper # - "noise_level_img": from [3, 25], set 3 for clean image, try 15 for very noisy LR images # - "k" (or "kernel_width"): blur kernel is very important!!! kernel_width from [0.6, 3.0] # to get the best performance. # ---------------------------------------------------------------------------------- """ ############################################################################## testset_name = 'Set3C' # set test set, 'set5' | 'srbsd68' noise_level_img = 3 # set noise level of image, from [3, 25], set 3 for clean image model_name = 'drunet_color' # 'ircnn_color' # set denoiser, | 'drunet_color' | 'ircnn_gray' | 'drunet_gray' | 'ircnn_color' sf = 2 # set scale factor, 1, 2, 3, 4 iter_num = 24 # set number of iterations, default: 24 for SISR # -------------------------------- # set blur kernel # -------------------------------- kernel_width_default_x1234 = [ 0.6, 0.9, 1.7, 2.2 ] # Gaussian kernel widths for x1, x2, x3, x4 noise_level_model = noise_level_img / 255. # noise level of model kernel_width = kernel_width_default_x1234[sf - 1] """ # set your own kernel width !!!!!!!!!! """ # kernel_width = 1.0 k = utils_deblur.fspecial('gaussian', 25, kernel_width) k = sr.shift_pixel(k, sf) # shift the kernel k /= np.sum(k) ############################################################################## show_img = False util.surf(k) if show_img else None x8 = True # default: False, x8 to boost performance modelSigma1 = 49 # set sigma_1, default: 49 modelSigma2 = max(sf, noise_level_model * 255.) classical_degradation = True # set classical degradation or bicubic degradation task_current = 'sr' # 'sr' for super-resolution n_channels = 1 if 'gray' in model_name else 3 # fixed model_zoo = 'model_zoo' # fixed testsets = 'testsets' # fixed results = 'results' # fixed result_name = testset_name + '_realapplications_' + task_current + '_' + model_name model_path = os.path.join(model_zoo, model_name + '.pth') device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') torch.cuda.empty_cache() # ---------------------------------------- # L_path, E_path, H_path # ---------------------------------------- L_path = os.path.join(testsets, testset_name) # L_path, for Low-quality images E_path = os.path.join(results, result_name) # E_path, for Estimated images util.mkdir(E_path) logger_name = result_name utils_logger.logger_info(logger_name, log_path=os.path.join(E_path, logger_name + '.log')) logger = logging.getLogger(logger_name) # ---------------------------------------- # load model # ---------------------------------------- if 'drunet' in model_name: from models.network_unet import UNetRes as net model = net(in_nc=n_channels + 1, out_nc=n_channels, nc=[64, 128, 256, 512], nb=4, act_mode='R', downsample_mode="strideconv", upsample_mode="convtranspose") model.load_state_dict(torch.load(model_path), strict=True) model.eval() for _, v in model.named_parameters(): v.requires_grad = False model = model.to(device) elif 'ircnn' in model_name: from models.network_dncnn import IRCNN as net model = net(in_nc=n_channels, out_nc=n_channels, nc=64) model25 = torch.load(model_path) former_idx = 0 logger.info('model_name:{}, image sigma:{:.3f}, model sigma:{:.3f}'.format( model_name, noise_level_img, noise_level_model)) logger.info('Model path: {:s}'.format(model_path)) logger.info(L_path) L_paths = util.get_image_paths(L_path) for idx, img in enumerate(L_paths): # -------------------------------- # (1) get img_L # -------------------------------- logger.info('Model path: {:s} Image: {:s}'.format(model_path, img)) img_name, ext = os.path.splitext(os.path.basename(img)) img_L = util.imread_uint(img, n_channels=n_channels) img_L = util.uint2single(img_L) img_L = util.modcrop(img_L, 8) # modcrop # -------------------------------- # (2) get rhos and sigmas # -------------------------------- rhos, sigmas = pnp.get_rho_sigma(sigma=max(0.255 / 255., noise_level_model), iter_num=iter_num, modelSigma1=modelSigma1, modelSigma2=modelSigma2, w=1) rhos, sigmas = torch.tensor(rhos).to(device), torch.tensor(sigmas).to( device) # -------------------------------- # (3) initialize x, and pre-calculation # -------------------------------- x = cv2.resize(img_L, (img_L.shape[1] * sf, img_L.shape[0] * sf), interpolation=cv2.INTER_CUBIC) if np.ndim(x) == 2: x = x[..., None] if classical_degradation: x = sr.shift_pixel(x, sf) x = util.single2tensor4(x).to(device) img_L_tensor, k_tensor = util.single2tensor4( img_L), util.single2tensor4(np.expand_dims(k, 2)) [k_tensor, img_L_tensor] = util.todevice([k_tensor, img_L_tensor], device) FB, FBC, F2B, FBFy = sr.pre_calculate(img_L_tensor, k_tensor, sf) # -------------------------------- # (4) main iterations # -------------------------------- for i in range(iter_num): print('Iter: {} / {}'.format(i, iter_num)) # -------------------------------- # step 1, FFT # -------------------------------- tau = rhos[i].float().repeat(1, 1, 1, 1) x = sr.data_solution(x, FB, FBC, F2B, FBFy, tau, sf) if 'ircnn' in model_name: current_idx = np.int( np.ceil(sigmas[i].cpu().numpy() * 255. / 2.) - 1) if current_idx != former_idx: model.load_state_dict(model25[str(current_idx)], strict=True) model.eval() for _, v in model.named_parameters(): v.requires_grad = False model = model.to(device) former_idx = current_idx # -------------------------------- # step 2, denoiser # -------------------------------- if x8: x = util.augment_img_tensor4(x, i % 8) if 'drunet' in model_name: x = torch.cat( (x, sigmas[i].repeat(1, 1, x.shape[2], x.shape[3])), dim=1) x = utils_model.test_mode(model, x, mode=2, refield=64, min_size=256, modulo=16) elif 'ircnn' in model_name: x = model(x) if x8: if i % 8 == 3 or i % 8 == 5: x = util.augment_img_tensor4(x, 8 - i % 8) else: x = util.augment_img_tensor4(x, i % 8) # -------------------------------- # (3) img_E # -------------------------------- img_E = util.tensor2uint(x) util.imsave( img_E, os.path.join(E_path, img_name + '_x' + str(sf) + '_' + model_name + '.png'))
def main(): # ---------------------------------------- # load kernels # ---------------------------------------- PSF_grid = np.load('./data/AC254-075-A-ML-Zemax(ZMX).npz')['PSF'] PSF_grid = PSF_grid.astype(np.float32) gx, gy = PSF_grid.shape[:2] k_tensor = [] for yy in range(gy): for xx in range(gx): PSF_grid[xx, yy] = PSF_grid[xx, yy] / np.sum(PSF_grid[xx, yy], axis=(0, 1)) k_tensor.append(util.single2tensor4(PSF_grid[xx, yy])) k_tensor = torch.cat(k_tensor, dim=0) inv_weight = util_deblur.get_inv_spatial_weight(k_tensor) # ---------------------------------------- # load model # ---------------------------------------- device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') model = net(n_iter=8, h_nc=64, in_nc=4, out_nc=3, nc=[64, 128, 256, 512], nb=2, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") model.proj.load_state_dict(torch.load('./data/usrnet_pretrain.pth'), strict=True) model.train() for _, v in model.named_parameters(): v.requires_grad = True model = model.to(device) # ---------------------------------------- # load training data # ---------------------------------------- imgs = glob.glob('./DIV2K_train/*.png', recursive=True) imgs.sort() # ---------------------------------------- # positional lambda\mu for HQS # ---------------------------------------- stage = 8 ab_buffer = np.ones((gx, gy, 2 * stage, 3), dtype=np.float32) * 0.1 #ab_buffer[:,:,0,:] = 0.01 ab = torch.tensor(ab_buffer, device=device, requires_grad=True) # ---------------------------------------- # build optimizer # ---------------------------------------- params = [] params += [{"params": [ab], "lr": 0.0005}] for key, value in model.named_parameters(): params += [{"params": [value], "lr": 0.0001}] optimizer = torch.optim.Adam(params, lr=0.0001, betas=(0.9, 0.999)) scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=1000, gamma=0.9) patch_size = [128, 128] expand = PSF_grid.shape[2] // 2 patch_num = [2, 2] global_iter = 0 running = True while running: #alpha.beta img_idx = np.random.randint(len(imgs)) img = imgs[img_idx] img_H = cv2.imread(img) w, h = img_H.shape[:2] #focus on the edges mode = np.random.randint(5) px_start = np.random.randint(0, gx - patch_num[0] + 1) py_start = np.random.randint(0, gy - patch_num[1] + 1) if mode == 0: px_start = 0 if mode == 1: px_start = gx - patch_num[0] if mode == 2: py_start = 0 if mode == 3: py_start = gy - patch_num[1] x_start = np.random.randint( 0, w - patch_size[0] * patch_num[0] - expand * 2 + 1) y_start = np.random.randint( 0, h - patch_size[1] * patch_num[1] - expand * 2 + 1) PSF_patch = PSF_grid[px_start:px_start + patch_num[0], py_start:py_start + patch_num[1]] patch_H = img_H[x_start:x_start+patch_size[0]*patch_num[0]+expand*2,\ y_start:y_start+patch_size[1]*patch_num[1]+expand*2] patch_L = util_deblur.blockConv2d(patch_H, PSF_patch, expand) block_expand = max(patch_size[0] // 8, expand) patch_L_wrap = util_deblur.wrap_boundary_liu( patch_L, (patch_size[0] * patch_num[0] + block_expand * 2, patch_size[1] * patch_num[1] + block_expand * 2)) patch_L_wrap = np.hstack( (patch_L_wrap[:, -block_expand:, :], patch_L_wrap[:, :patch_size[1] * patch_num[1] + block_expand, :])) patch_L_wrap = np.vstack( (patch_L_wrap[-block_expand:, :, :], patch_L_wrap[:patch_size[0] * patch_num[0] + block_expand, :, :])) x = util.uint2single(patch_L_wrap) x = util.single2tensor4(x) x_gt = util.uint2single(patch_H[expand:-expand, expand:-expand]) x_gt = util.single2tensor4(x_gt) inv_weight_patch = torch.ones_like(x_gt) k_local = [] for h_ in range(patch_num[1]): for w_ in range(patch_num[0]): inv_weight_patch[0, 0, w_ * patch_size[0]:(w_ + 1) * patch_size[0], h_ * patch_size[1]:(h_ + 1) * patch_size[1]] = inv_weight[w_ + h_ * patch_num[0], 0] inv_weight_patch[0, 1, w_ * patch_size[0]:(w_ + 1) * patch_size[0], h_ * patch_size[1]:(h_ + 1) * patch_size[1]] = inv_weight[w_ + h_ * patch_num[0], 1] inv_weight_patch[0, 2, w_ * patch_size[0]:(w_ + 1) * patch_size[0], h_ * patch_size[1]:(h_ + 1) * patch_size[1]] = inv_weight[w_ + h_ * patch_num[0], 2] k_local.append(k_tensor[w_ + h_ * patch_num[0]:w_ + h_ * patch_num[0] + 1]) k = torch.cat(k_local, dim=0) [x, x_gt, k, inv_weight_patch ] = [el.to(device) for el in [x, x_gt, k, inv_weight_patch]] ab_patch = F.softplus(ab[px_start:px_start + patch_num[0], py_start:py_start + patch_num[1]]) cd = [] for h_ in range(patch_num[1]): for w_ in range(patch_num[0]): cd.append(ab_patch[w_:w_ + 1, h_]) cd = torch.cat(cd, dim=0) x_E = model.forward_patchwise(x, k, cd, patch_num, patch_size) predict = x_E[...,block_expand:block_expand+patch_size[0]*patch_num[0],\ block_expand:block_expand+patch_size[1]*patch_num[1]] loss = F.l1_loss(predict.div(inv_weight_patch), x_gt.div(inv_weight_patch)) optimizer.zero_grad() loss.backward() optimizer.step() scheduler.step() print('iter:{},loss {}'.format(global_iter + 1, loss.item())) patch_L = patch_L_wrap.astype(np.uint8) patch_E = util.tensor2uint(x_E)[block_expand:-block_expand, block_expand:-block_expand] show = np.hstack((patch_H[expand:-expand, expand:-expand], patch_L[block_expand:-block_expand, block_expand:-block_expand], patch_E)) cv2.imshow('HL', show) key = cv2.waitKey(1) global_iter += 1 #change the save period if global_iter % 100 == 0: ab_numpy = ab.detach().cpu().numpy().flatten() torch.save( model.state_dict(), './ZEMAX_model/usrnet_ZEMAX_iter{}.pth'.format(global_iter)) np.savetxt('./ZEMAX_model/ab_ZEMAX_iter{}.txt'.format(global_iter), ab_numpy) if key == ord('q'): running = False break ab_numpy = ab.detach().cpu().numpy().flatten() torch.save(model.state_dict(), './ZEMAX_model/usrnet_ZEMAX.pth') np.savetxt('./ZEMAX_model/ab_ZEMAX.txt', ab_numpy)
def main(): # ---------------------------------------- # Preparation # ---------------------------------------- noise_level_img = 0 # default: 0, noise level for LR image noise_level_model = noise_level_img # noise level for model model_name = 'srmdnf_x4' # 'srmd_x2' | 'srmd_x3' | 'srmd_x4' | 'srmdnf_x2' | 'srmdnf_x3' | 'srmdnf_x4' testset_name = 'set5' # test set, 'set5' | 'srbsd68' sf = [int(s) for s in re.findall(r'\d+', model_name)][0] # scale factor x8 = False # default: False, x8 to boost performance need_degradation = True # default: True, use degradation model to generate LR image show_img = False # default: False srmd_pca_path = os.path.join('kernels', 'srmd_pca_matlab.mat') task_current = 'sr' # 'dn' for denoising | 'sr' for super-resolution n_channels = 3 # fixed in_nc = 18 if 'nf' in model_name else 19 nc = 128 # fixed, number of channels nb = 12 # fixed, number of conv layers model_pool = 'model_zoo' # fixed testsets = 'testsets' # fixed results = 'results' # fixed result_name = testset_name + '_' + model_name border = sf if task_current == 'sr' else 0 # shave boader to calculate PSNR and SSIM model_path = os.path.join(model_pool, model_name + '.pth') # ---------------------------------------- # L_path, E_path, H_path # ---------------------------------------- L_path = os.path.join(testsets, testset_name) # L_path, for Low-quality images H_path = L_path # H_path, for High-quality images E_path = os.path.join(results, result_name) # E_path, for Estimated images util.mkdir(E_path) if H_path == L_path: need_degradation = True logger_name = result_name utils_logger.logger_info(logger_name, log_path=os.path.join(E_path, logger_name + '.log')) logger = logging.getLogger(logger_name) need_H = True if H_path is not None else False device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') # ---------------------------------------- # load model # ---------------------------------------- from models.network_srmd import SRMD as net model = net(in_nc=in_nc, out_nc=n_channels, nc=nc, nb=nb, upscale=sf, act_mode='R', upsample_mode='pixelshuffle') model.load_state_dict(torch.load(model_path), strict=False) model.eval() for k, v in model.named_parameters(): v.requires_grad = False model = model.to(device) logger.info('Model path: {:s}'.format(model_path)) number_parameters = sum(map(lambda x: x.numel(), model.parameters())) logger.info('Params number: {}'.format(number_parameters)) test_results = OrderedDict() test_results['psnr'] = [] test_results['ssim'] = [] test_results['psnr_y'] = [] test_results['ssim_y'] = [] logger.info('model_name:{}, model sigma:{}, image sigma:{}'.format( model_name, noise_level_img, noise_level_model)) logger.info(L_path) L_paths = util.get_image_paths(L_path) H_paths = util.get_image_paths(H_path) if need_H else None # ---------------------------------------- # kernel and PCA reduced feature # ---------------------------------------- # kernel = sr.anisotropic_Gaussian(ksize=15, theta=np.pi, l1=4, l2=4) kernel = utils_deblur.fspecial('gaussian', 15, 0.01) # Gaussian kernel, delta kernel 0.01 P = loadmat(srmd_pca_path)['P'] degradation_vector = np.dot(P, np.reshape(kernel, (-1), order="F")) if 'nf' not in model_name: # noise-free SR degradation_vector = np.append(degradation_vector, noise_level_model / 255.) degradation_vector = torch.from_numpy(degradation_vector).view( 1, -1, 1, 1).float() for idx, img in enumerate(L_paths): # ------------------------------------ # (1) img_L # ------------------------------------ img_name, ext = os.path.splitext(os.path.basename(img)) # logger.info('{:->4d}--> {:>10s}'.format(idx+1, img_name+ext)) img_L = util.imread_uint(img, n_channels=n_channels) img_L = util.uint2single(img_L) # degradation process, blur + bicubic downsampling + Gaussian noise if need_degradation: img_L = util.modcrop(img_L, sf) img_L = sr.srmd_degradation( img_L, kernel, sf ) # equivalent to bicubic degradation if kernel is a delta kernel np.random.seed(seed=0) # for reproducibility img_L += np.random.normal(0, noise_level_img / 255., img_L.shape) util.imshow(util.single2uint(img_L), title='LR image with noise level {}'.format( noise_level_img)) if show_img else None img_L = util.single2tensor4(img_L) degradation_map = degradation_vector.repeat(1, 1, img_L.size(-2), img_L.size(-1)) img_L = torch.cat((img_L, degradation_map), dim=1) img_L = img_L.to(device) # ------------------------------------ # (2) img_E # ------------------------------------ if not x8: img_E = model(img_L) else: img_E = utils_model.test_mode(model, img_L, mode=3, sf=sf) img_E = util.tensor2uint(img_E) if need_H: # -------------------------------- # (3) img_H # -------------------------------- img_H = util.imread_uint(H_paths[idx], n_channels=n_channels) img_H = img_H.squeeze() img_H = util.modcrop(img_H, sf) # -------------------------------- # PSNR and SSIM # -------------------------------- psnr = util.calculate_psnr(img_E, img_H, border=border) ssim = util.calculate_ssim(img_E, img_H, border=border) test_results['psnr'].append(psnr) test_results['ssim'].append(ssim) logger.info('{:s} - PSNR: {:.2f} dB; SSIM: {:.4f}.'.format( img_name + ext, psnr, ssim)) util.imshow(np.concatenate([img_E, img_H], axis=1), title='Recovered / Ground-truth') if show_img else None if np.ndim(img_H) == 3: # RGB image img_E_y = util.rgb2ycbcr(img_E, only_y=True) img_H_y = util.rgb2ycbcr(img_H, only_y=True) psnr_y = util.calculate_psnr(img_E_y, img_H_y, border=border) ssim_y = util.calculate_ssim(img_E_y, img_H_y, border=border) test_results['psnr_y'].append(psnr_y) test_results['ssim_y'].append(ssim_y) # ------------------------------------ # save results # ------------------------------------ util.imsave(img_E, os.path.join(E_path, img_name + '.png')) if need_H: ave_psnr = sum(test_results['psnr']) / len(test_results['psnr']) ave_ssim = sum(test_results['ssim']) / len(test_results['ssim']) logger.info( 'Average PSNR/SSIM(RGB) - {} - x{} --PSNR: {:.2f} dB; SSIM: {:.4f}' .format(result_name, sf, ave_psnr, ave_ssim)) if np.ndim(img_H) == 3: ave_psnr_y = sum(test_results['psnr_y']) / len( test_results['psnr_y']) ave_ssim_y = sum(test_results['ssim_y']) / len( test_results['ssim_y']) logger.info( 'Average PSNR/SSIM( Y ) - {} - x{} - PSNR: {:.2f} dB; SSIM: {:.4f}' .format(result_name, sf, ave_psnr_y, ave_ssim_y))
def main(): #0. global config sf = 4 stage = 8 patch_size = [32, 32] patch_num = [2, 2] #1. local PSF all_PSFs = load_kernels('./data') #2. load model device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') model = net(n_iter=8, h_nc=64, in_nc=4, out_nc=3, nc=[64, 128, 256, 512], nb=2, sf=sf, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") model.load_state_dict(torch.load('./data/uabcnet_final.pth'), strict=True) model.train() for _, v in model.named_parameters(): v.requires_grad = True model = model.to(device) #3. set up discriminator model_D = gan.PatchDiscriminator(5) model_D = model_D.to(device) gan_loss = gan.GANLoss(mode='lsgan') gan_loss = gan_loss.to(device) fake_images = ImagePool(16) #positional lambda, mu for HQS. ab_buffer = np.zeros((patch_num[0], patch_num[1], 2 * stage, 3)) ab_buffer[:, :, ::2, :] = 0.01 ab_buffer[:, :, 1::2, :] = 0.1 ab = torch.tensor(ab_buffer, dtype=torch.float32, device=device, requires_grad=True) params = [] params += [{"params": [ab], "lr": 5e-4}] for key, value in model.named_parameters(): params += [{"params": [value], "lr": 1e-5}] # params_D = [] params_D += list(model_D.parameters()) optimizer = torch.optim.Adam(params, lr=1e-4, betas=(0.9, 0.999)) optimizer_D = torch.optim.Adam(params_D, lr=1e-4, betas=(0.9, 0.999)) scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=1000, gamma=0.9) #3.load training data imgs_H = glob.glob('/home/xiu/databag/deblur/images/DIV2K_train/*.png', recursive=True) imgs_H.sort() global_iter = 0 N_maxiter = 200000 PSF_grid = draw_random_kernel(all_PSFs) for i in range(N_maxiter): t0 = time.time() img_idx = np.random.randint(len(imgs_H)) img_H = cv2.imread(imgs_H[img_idx]) #draw random kernel patch_L, patch_H, patch_psf = draw_training_pair( img_H, PSF_grid, sf, patch_num, patch_size) t_data = time.time() - t0 x = util.uint2single(patch_L) x = util.single2tensor4(x) x_gt = util.uint2single(patch_H) x_gt = util.single2tensor4(x_gt) k_local = [] for h_ in range(patch_num[1]): for w_ in range(patch_num[0]): k_local.append(util.single2tensor4(patch_psf[w_, h_])) k = torch.cat(k_local, dim=0) [x, x_gt, k] = [el.to(device) for el in [x, x_gt, k]] ab_patch = F.softplus(ab) ab_patch_v = [] for h_ in range(patch_num[1]): for w_ in range(patch_num[0]): ab_patch_v.append(ab_patch[w_:w_ + 1, h_]) ab_patch_v = torch.cat(ab_patch_v, dim=0) x_E = model.forward_patchwise_SR(x, k, ab_patch_v, patch_num, [patch_size[0], patch_size[1]], sf) loss_l1 = F.l1_loss(x_E, x_gt) loss_gan = gan_loss(model_D(x_E), True) loss = loss_l1 + loss_gan optimizer.zero_grad() loss.backward() optimizer.step() pred_real = model_D(x_gt) loss_D_real = gan_loss(pred_real, True) fake = fake_images.query(x_E) pred_fake = model_D(fake.detach()) loss_D_fake = gan_loss(pred_fake, False) loss_D = (loss_D_fake + loss_D_real) * 0.5 optimizer_D.zero_grad() loss_D.backward() optimizer_D.step() scheduler.step() t_iter = time.time() - t0 - t_data print('[iter:{}] loss:{:.4f}, data_time:{:.2f}s, net_time:{:.2f}s'. format(global_iter + 1, loss.item(), t_data, t_iter)) patch_L = cv2.resize(patch_L, dsize=None, fx=sf, fy=sf, interpolation=cv2.INTER_NEAREST) patch_E = util.tensor2uint((x_E)) show = np.hstack((patch_H, patch_L, patch_E)) cv2.imshow('H,L,E', show) key = cv2.waitKey(1) global_iter += 1 if key == ord('q'): break ab_numpy = ab.detach().cpu().numpy().flatten() torch.save(model.state_dict(), './data/uabcnet_finetune.pth') np.savetxt('./data/ab_finetune.txt', ab_numpy)
def __getitem__(self, index): # ------------------------------------ # get H image # ------------------------------------ H_path = self.paths_H[index] img_H = util.imread_uint(H_path, self.n_channels) img_H = util.uint2single(img_H) # ------------------------------------ # modcrop for SR # ------------------------------------ img_H = util.modcrop(img_H, self.sf) # ------------------------------------ # sythesize L image via matlab's bicubic # ------------------------------------ H, W, _ = img_H.shape img_L = util.imresize_np(img_H, 1 / self.sf, True) if self.opt['phase'] == 'train': """ # -------------------------------- # get L/H patch pairs # -------------------------------- """ H, W, C = img_L.shape # -------------------------------- # randomly crop L patch # -------------------------------- rnd_h = random.randint(0, max(0, H - self.L_size)) rnd_w = random.randint(0, max(0, W - self.L_size)) img_L = img_L[rnd_h:rnd_h + self.L_size, rnd_w:rnd_w + self.L_size, :] # -------------------------------- # crop corresponding H patch # -------------------------------- rnd_h_H, rnd_w_H = int(rnd_h * self.sf), int(rnd_w * self.sf) img_H = img_H[rnd_h_H:rnd_h_H + self.patch_size, rnd_w_H:rnd_w_H + self.patch_size, :] # -------------------------------- # augmentation - flip and/or rotate # -------------------------------- mode = np.random.randint(0, 8) img_L, img_H = util.augment_img( img_L, mode=mode), util.augment_img(img_H, mode=mode) # -------------------------------- # get patch pairs # -------------------------------- img_H, img_L = util.single2tensor3(img_H), util.single2tensor3( img_L) # -------------------------------- # select noise level and get Gaussian noise # -------------------------------- if random.random() < 0.1: noise_level = torch.zeros(1).float() else: noise_level = torch.FloatTensor([ np.random.uniform(self.sigma_min, self.sigma_max) ]) / 255.0 # noise_level = torch.rand(1)*50/255.0 # noise_level = torch.min(torch.from_numpy(np.float32([7*np.random.chisquare(2.5)/255.0])),torch.Tensor([50./255.])) else: img_H, img_L = util.single2tensor3(img_H), util.single2tensor3( img_L) noise_level = torch.FloatTensor([self.sigma_test]) # ------------------------------------ # add noise # ------------------------------------ noise = torch.randn(img_L.size()).mul_(noise_level).float() img_L.add_(noise) # ------------------------------------ # get noise level map M # ------------------------------------ M_vector = noise_level.unsqueeze(1).unsqueeze(1) M = M_vector.repeat(1, img_L.size()[-2], img_L.size()[-1]) """ # ------------------------------------- # concat L and noise level map M # ------------------------------------- """ img_L = torch.cat((img_L, M), 0) L_path = H_path return {'L': img_L, 'H': img_H, 'L_path': L_path, 'H_path': H_path}
def main(): #0. global config #scale factor sf = 4 stage = 5 patch_size = [32,32] patch_num = [2,2] #1. local PSF #shape: gx,gy,kw,kw,3 all_PSFs = load_kernels('./data') #2. local model device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') model = net(n_iter=5, h_nc=64, in_nc=4, out_nc=3, nc=[64, 128, 256, 512], nb=2,sf=sf, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") model.load_state_dict(torch.load('./logs/uabcnet_final.pth'),strict=True) model.train() for _, v in model.named_parameters(): v.requires_grad = True model = model.to(device) #positional lambda, mu for HQS, set as free trainable parameters here. #ab_buffer = np.loadtxt('./data/ab.txt').reshape((patch_num[0],patch_num[1],2*stage,3)).astype(np.float32) ab_pretrain = np.loadtxt('./logs/ab_pretrain.txt').reshape((1,1,2*stage,3)).astype(np.float32) ab_buffer = np.ones((patch_num[0],patch_num[1],2*stage,3),dtype=np.float32) for xx in range(patch_num[0]): for yy in range(patch_num[1]): ab_buffer[xx,yy] = ab_pretrain[0,0] ab = torch.tensor(ab_buffer,device=device,requires_grad=True) params = [] params += [{"params":[ab],"lr":0.0001}] for key,value in model.named_parameters(): params += [{"params":[value],"lr":1e-6}] optimizer = torch.optim.Adam(params,lr=0.0001,betas=(0.9,0.999)) scheduler = torch.optim.lr_scheduler.StepLR(optimizer,step_size=1000,gamma=0.9) #3.load training data imgs_H = glob.glob('./DIV2K_train/*.png',recursive=True) imgs_H.sort() global_iter = 0 all_PSNR = [] N_maxiter = 4000 PSF_grid = draw_random_kernel(all_PSFs,patch_num) #def get_train_pairs() for i in range(N_maxiter): t0 = time.time() #draw random image. img_idx = np.random.randint(len(imgs_H)) img_H = cv2.imread(imgs_H[img_idx]) patch_L,patch_H,patch_psf = draw_training_pair(img_H,PSF_grid,sf,patch_num,patch_size) #b. with img_L #patch_L, patch_H, patch_psf,px_start, py_start,block_expand = draw_training_pair(img_H, PSF_grid, sf, patch_num, patch_size, img_L) t_data = time.time()-t0 x = util.uint2single(patch_L) x = util.single2tensor4(x) x_gt = util.uint2single(patch_H) x_gt = util.single2tensor4(x_gt) k_local = [] for h_ in range(patch_num[1]): for w_ in range(patch_num[0]): k_local.append(util.single2tensor4(patch_psf[w_,h_])) k = torch.cat(k_local,dim=0) [x,x_gt,k] = [el.to(device) for el in [x,x_gt,k]] ab_patch = F.softplus(ab) ab_patch_v = [] for h_ in range(patch_num[1]): for w_ in range(patch_num[0]): ab_patch_v.append(ab_patch[w_:w_+1,h_]) ab_patch_v = torch.cat(ab_patch_v,dim=0) x_E = model.forward_patchwise_SR(x,k,ab_patch_v,patch_num,[patch_size[0],patch_size[1]],sf) loss = F.l1_loss(x_E,x_gt) optimizer.zero_grad() loss.backward() optimizer.step() scheduler.step() t_iter = time.time() - t0 - t_data print('[iter:{}] loss:{:.4f}, data_time:{:.2f}s, net_time:{:.2f}s'.format(global_iter+1,loss.item(),t_data,t_iter)) patch_L = cv2.resize(patch_L,dsize=None,fx=sf,fy=sf,interpolation=cv2.INTER_NEAREST) #patch_L = patch_L[block_expand*sf:-block_expand*sf,block_expand*sf:-block_expand*sf] patch_E = util.tensor2uint((x_E)) show = np.hstack((patch_H,patch_L,patch_E)) cv2.imshow('H,L,E',show) key = cv2.waitKey(1) global_iter+= 1 if i % 1000 ==0: cv2.imwrite(os.path.join('./result', 'test' , 'resultE-{:04d}.png'.format(i + 1)), patch_E) cv2.imwrite(os.path.join('./result', 'test', 'resultL-{:04d}.png'.format(i + 1)), patch_L) cv2.imwrite(os.path.join('./result', 'test', 'resultH-{:04d}.png'.format(i + 1)), patch_H) # if key==ord('q'): # break # if key==ord('s'): # ab_numpy = ab.detach().cpu().numpy().flatten() # np.savetxt('./data/ab.txt',ab_numpy) ab_numpy = ab.detach().cpu().numpy().flatten() torch.save(model.state_dict(),'./data/uabcnet_finetune.pth') np.savetxt('./data/ab_finetune.txt',ab_numpy)
def main(): # ---------------------------------------- # Preparation # ---------------------------------------- noise_level_img = 0/255.0 # set AWGN noise level for LR image, default: 0, noise_level_model = noise_level_img # setnoise level of model, default 0 model_name = 'drunet_color' # set denoiser, | 'drunet_color' | 'ircnn_gray' | 'drunet_gray' | 'ircnn_color' testset_name = 'srbsd68' # set test set, 'set5' | 'srbsd68' x8 = True # default: False, x8 to boost performance test_sf = [2] # set scale factor, default: [2, 3, 4], [2], [3], [4] iter_num = 24 # set number of iterations, default: 24 for SISR modelSigma1 = 49 # set sigma_1, default: 49 classical_degradation = True # set classical degradation or bicubic degradation show_img = False # default: False save_L = True # save LR image save_E = True # save estimated image save_LEH = False # save zoomed LR, E and H images task_current = 'sr' # 'sr' for super-resolution n_channels = 1 if 'gray' in model_name else 3 # fixed model_zoo = 'model_zoo' # fixed testsets = 'testsets' # fixed results = 'results' # fixed result_name = testset_name + '_' + task_current + '_' + model_name model_path = os.path.join(model_zoo, model_name+'.pth') device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') torch.cuda.empty_cache() # ---------------------------------------- # L_path, E_path, H_path # ---------------------------------------- L_path = os.path.join(testsets, testset_name) # L_path, for Low-quality images E_path = os.path.join(results, result_name) # E_path, for Estimated images util.mkdir(E_path) logger_name = result_name utils_logger.logger_info(logger_name, log_path=os.path.join(E_path, logger_name+'.log')) logger = logging.getLogger(logger_name) # ---------------------------------------- # load model # ---------------------------------------- if 'drunet' in model_name: from models.network_unet import UNetRes as net model = net(in_nc=n_channels+1, out_nc=n_channels, nc=[64, 128, 256, 512], nb=4, act_mode='R', downsample_mode="strideconv", upsample_mode="convtranspose") model.load_state_dict(torch.load(model_path), strict=True) model.eval() for _, v in model.named_parameters(): v.requires_grad = False model = model.to(device) elif 'ircnn' in model_name: from models.network_dncnn import IRCNN as net model = net(in_nc=n_channels, out_nc=n_channels, nc=64) model25 = torch.load(model_path) former_idx = 0 logger.info('model_name:{}, image sigma:{:.3f}, model sigma:{:.3f}'.format(model_name, noise_level_img, noise_level_model)) logger.info('Model path: {:s}'.format(model_path)) logger.info(L_path) L_paths = util.get_image_paths(L_path) # -------------------------------- # load kernel # -------------------------------- # kernels = hdf5storage.loadmat(os.path.join('kernels', 'Levin09.mat'))['kernels'] if classical_degradation: kernels = hdf5storage.loadmat(os.path.join('kernels', 'kernels_12.mat'))['kernels'] else: kernels = hdf5storage.loadmat(os.path.join('kernels', 'kernel_bicubicx234.mat'))['kernels'] test_results_ave = OrderedDict() test_results_ave['psnr_sf_k'] = [] test_results_ave['psnr_y_sf_k'] = [] for sf in test_sf: border = sf modelSigma2 = max(sf, noise_level_model*255.) k_num = 8 if classical_degradation else 1 for k_index in range(k_num): logger.info('--------- sf:{:>1d} --k:{:>2d} ---------'.format(sf, k_index)) test_results = OrderedDict() test_results['psnr'] = [] test_results['psnr_y'] = [] if not classical_degradation: # for bicubic degradation k_index = sf-2 k = kernels[0, k_index].astype(np.float64) util.surf(k) if show_img else None for idx, img in enumerate(L_paths): # -------------------------------- # (1) get img_L # -------------------------------- img_name, ext = os.path.splitext(os.path.basename(img)) img_H = util.imread_uint(img, n_channels=n_channels) img_H = util.modcrop(img_H, sf) # modcrop if classical_degradation: img_L = sr.classical_degradation(img_H, k, sf) util.imshow(img_L) if show_img else None img_L = util.uint2single(img_L) else: img_L = util.imresize_np(util.uint2single(img_H), 1/sf) np.random.seed(seed=0) # for reproducibility img_L += np.random.normal(0, noise_level_img, img_L.shape) # add AWGN # -------------------------------- # (2) get rhos and sigmas # -------------------------------- rhos, sigmas = pnp.get_rho_sigma(sigma=max(0.255/255., noise_level_model), iter_num=iter_num, modelSigma1=modelSigma1, modelSigma2=modelSigma2, w=1) rhos, sigmas = torch.tensor(rhos).to(device), torch.tensor(sigmas).to(device) # -------------------------------- # (3) initialize x, and pre-calculation # -------------------------------- x = cv2.resize(img_L, (img_L.shape[1]*sf, img_L.shape[0]*sf), interpolation=cv2.INTER_CUBIC) if np.ndim(x)==2: x = x[..., None] if classical_degradation: x = sr.shift_pixel(x, sf) x = util.single2tensor4(x).to(device) img_L_tensor, k_tensor = util.single2tensor4(img_L), util.single2tensor4(np.expand_dims(k, 2)) [k_tensor, img_L_tensor] = util.todevice([k_tensor, img_L_tensor], device) FB, FBC, F2B, FBFy = sr.pre_calculate(img_L_tensor, k_tensor, sf) # -------------------------------- # (4) main iterations # -------------------------------- for i in range(iter_num): # -------------------------------- # step 1, FFT # -------------------------------- tau = rhos[i].float().repeat(1, 1, 1, 1) x = sr.data_solution(x.float(), FB, FBC, F2B, FBFy, tau, sf) if 'ircnn' in model_name: current_idx = np.int(np.ceil(sigmas[i].cpu().numpy()*255./2.)-1) if current_idx != former_idx: model.load_state_dict(model25[str(current_idx)], strict=True) model.eval() for _, v in model.named_parameters(): v.requires_grad = False model = model.to(device) former_idx = current_idx # -------------------------------- # step 2, denoiser # -------------------------------- if x8: x = util.augment_img_tensor4(x, i % 8) if 'drunet' in model_name: x = torch.cat((x, sigmas[i].float().repeat(1, 1, x.shape[2], x.shape[3])), dim=1) x = utils_model.test_mode(model, x, mode=2, refield=32, min_size=256, modulo=16) elif 'ircnn' in model_name: x = model(x) if x8: if i % 8 == 3 or i % 8 == 5: x = util.augment_img_tensor4(x, 8 - i % 8) else: x = util.augment_img_tensor4(x, i % 8) # -------------------------------- # (3) img_E # -------------------------------- img_E = util.tensor2uint(x) if save_E: util.imsave(img_E, os.path.join(E_path, img_name+'_x'+str(sf)+'_k'+str(k_index)+'_'+model_name+'.png')) if n_channels == 1: img_H = img_H.squeeze() # -------------------------------- # (4) img_LEH # -------------------------------- img_L = util.single2uint(img_L).squeeze() if save_LEH: k_v = k/np.max(k)*1.0 if n_channels==1: k_v = util.single2uint(k_v) else: k_v = util.single2uint(np.tile(k_v[..., np.newaxis], [1, 1, n_channels])) k_v = cv2.resize(k_v, (3*k_v.shape[1], 3*k_v.shape[0]), interpolation=cv2.INTER_NEAREST) img_I = cv2.resize(img_L, (sf*img_L.shape[1], sf*img_L.shape[0]), interpolation=cv2.INTER_NEAREST) img_I[:k_v.shape[0], -k_v.shape[1]:, ...] = k_v img_I[:img_L.shape[0], :img_L.shape[1], ...] = img_L util.imshow(np.concatenate([img_I, img_E, img_H], axis=1), title='LR / Recovered / Ground-truth') if show_img else None util.imsave(np.concatenate([img_I, img_E, img_H], axis=1), os.path.join(E_path, img_name+'_x'+str(sf)+'_k'+str(k_index)+'_LEH.png')) if save_L: util.imsave(img_L, os.path.join(E_path, img_name+'_x'+str(sf)+'_k'+str(k_index)+'_LR.png')) psnr = util.calculate_psnr(img_E, img_H, border=border) test_results['psnr'].append(psnr) logger.info('{:->4d}--> {:>10s} -- sf:{:>1d} --k:{:>2d} PSNR: {:.2f}dB'.format(idx+1, img_name+ext, sf, k_index, psnr)) if n_channels == 3: img_E_y = util.rgb2ycbcr(img_E, only_y=True) img_H_y = util.rgb2ycbcr(img_H, only_y=True) psnr_y = util.calculate_psnr(img_E_y, img_H_y, border=border) test_results['psnr_y'].append(psnr_y) # -------------------------------- # Average PSNR for all kernels # -------------------------------- ave_psnr_k = sum(test_results['psnr']) / len(test_results['psnr']) logger.info('------> Average PSNR(RGB) of ({}) scale factor: ({}), kernel: ({}) sigma: ({:.2f}): {:.2f} dB'.format(testset_name, sf, k_index, noise_level_model, ave_psnr_k)) test_results_ave['psnr_sf_k'].append(ave_psnr_k) if n_channels == 3: # RGB image ave_psnr_y_k = sum(test_results['psnr_y']) / len(test_results['psnr_y']) logger.info('------> Average PSNR(Y) of ({}) scale factor: ({}), kernel: ({}) sigma: ({:.2f}): {:.2f} dB'.format(testset_name, sf, k_index, noise_level_model, ave_psnr_y_k)) test_results_ave['psnr_y_sf_k'].append(ave_psnr_y_k) # --------------------------------------- # Average PSNR for all sf and kernels # --------------------------------------- ave_psnr_sf_k = sum(test_results_ave['psnr_sf_k']) / len(test_results_ave['psnr_sf_k']) logger.info('------> Average PSNR of ({}) {:.2f} dB'.format(testset_name, ave_psnr_sf_k)) if n_channels == 3: ave_psnr_y_sf_k = sum(test_results_ave['psnr_y_sf_k']) / len(test_results_ave['psnr_y_sf_k']) logger.info('------> Average PSNR of ({}) {:.2f} dB'.format(testset_name, ave_psnr_y_sf_k))
def __getitem__(self, index): # ------------------- # get H image # ------------------- H_path = self.paths_H[index] img_H = util.imread_uint(H_path, self.n_channels) L_path = H_path if self.opt['phase'] == 'train': # --------------------------- # 1) scale factor, ensure each batch only involves one scale factor # --------------------------- if self.count % self.opt['dataloader_batch_size'] == 0: # sf = random.choice([1,2,3,4]) self.sf = random.choice(self.scales) # self.count = 0 # optional self.count += 1 H, W, _ = img_H.shape # ---------------------------- # randomly crop the patch # ---------------------------- rnd_h = random.randint(0, max(0, H - self.patch_size)) rnd_w = random.randint(0, max(0, W - self.patch_size)) patch_H = img_H[rnd_h:rnd_h + self.patch_size, rnd_w:rnd_w + self.patch_size, :] # --------------------------- # augmentation - flip, rotate # --------------------------- mode = np.random.randint(0, 8) patch_H = util.augment_img(patch_H, mode=mode) # --------------------------- # 2) kernel # --------------------------- r_value = random.randint(0, 7) if r_value > 3: k = utils_deblur.blurkernel_synthesis(h=25) # motion blur else: sf_k = random.choice(self.scales) k = utils_sisr.gen_kernel(scale_factor=np.array( [sf_k, sf_k])) # Gaussian blur mode_k = random.randint(0, 7) k = util.augment_img(k, mode=mode_k) # --------------------------- # 3) noise level # --------------------------- if random.randint(0, 8) == 1: noise_level = 0 / 255.0 else: noise_level = np.random.randint(0, self.sigma_max) / 255.0 # --------------------------- # Low-quality image # --------------------------- img_L = ndimage.filters.convolve(patch_H, np.expand_dims(k, axis=2), mode='wrap') img_L = img_L[0::self.sf, 0::self.sf, ...] # add Gaussian noise img_L = util.uint2single(img_L) + np.random.normal( 0, noise_level, img_L.shape) img_H = patch_H else: k = self.kernels[0, 0].astype(np.float64) # validation kernel k /= np.sum(k) noise_level = 0. / 255.0 # validation noise level img_L = ndimage.filters.convolve(img_H, np.expand_dims(k, axis=2), mode='wrap') # blur img_L = img_L[0::self.sf_validation, 0::self.sf_validation, ...] # downsampling img_L = util.uint2single(img_L) + np.random.normal( 0, noise_level, img_L.shape) k = util.single2tensor3(np.expand_dims(np.float32(k), axis=2)) img_H, img_L = util.uint2tensor3(img_H), util.single2tensor3(img_L) noise_level = torch.FloatTensor([noise_level]).view([1, 1, 1]) return { 'L': img_L, 'H': img_H, 'k': k, 'sigma': noise_level, 'sf': self.sf, 'L_path': L_path, 'H_path': H_path }
def main(): # ---------------------------------------- # Preparation # ---------------------------------------- model_name = 'usrnet' # 'usrgan' | 'usrnet' | 'usrgan_tiny' | 'usrnet_tiny' testset_name = 'set_real' # test set, 'set_real' test_image = 'chip.png' # 'chip.png', 'comic.png' #test_image = 'comic.png' sf = 4 # scale factor, only from {1, 2, 3, 4} show_img = False # default: False save_E = True # save estimated image save_LE = True # save zoomed LR, Estimated images # ---------------------------------------- # set noise level and kernel # ---------------------------------------- if 'chip' in test_image: noise_level_img = 15 # noise level for LR image, 15 for chip kernel_width_default_x1234 = [0.6, 0.9, 1.7, 2.2] # Gaussian kernel widths for x1, x2, x3, x4 else: noise_level_img = 2 # noise level for LR image, 0.5~3 for clean images kernel_width_default_x1234 = [0.4, 0.7, 1.5, 2.0] # default Gaussian kernel widths of clean/sharp images for x1, x2, x3, x4 noise_level_model = noise_level_img/255. # noise level of model kernel_width = kernel_width_default_x1234[sf-1] # set your own kernel width # kernel_width = 2.2 k = utils_deblur.fspecial('gaussian', 25, kernel_width) k = sr.shift_pixel(k, sf) # shift the kernel k /= np.sum(k) util.surf(k) if show_img else None # scio.savemat('kernel_realapplication.mat', {'kernel':k}) # load approximated bicubic kernels #kernels = hdf5storage.loadmat(os.path.join('kernels', 'kernel_bicubicx234.mat'))['kernels'] # kernels = loadmat(os.path.join('kernels', 'kernel_bicubicx234.mat'))['kernels'] # kernel = kernels[0, sf-2].astype(np.float64) kernel = util.single2tensor4(k[..., np.newaxis]) n_channels = 1 if 'gray' in model_name else 3 # 3 for color image, 1 for grayscale image model_pool = 'model_zoo' # fixed testsets = 'testsets' # fixed results = 'results' # fixed result_name = testset_name + '_' + model_name model_path = os.path.join(model_pool, model_name+'.pth') # ---------------------------------------- # L_path, E_path # ---------------------------------------- L_path = os.path.join(testsets, testset_name) # L_path, fixed, for Low-quality images E_path = os.path.join(results, result_name) # E_path, fixed, for Estimated images util.mkdir(E_path) logger_name = result_name utils_logger.logger_info(logger_name, log_path=os.path.join(E_path, logger_name+'.log')) logger = logging.getLogger(logger_name) device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') # ---------------------------------------- # load model # ---------------------------------------- if 'tiny' in model_name: model = net(n_iter=6, h_nc=32, in_nc=4, out_nc=3, nc=[16, 32, 64, 64], nb=2, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") else: model = net(n_iter=8, h_nc=64, in_nc=4, out_nc=3, nc=[64, 128, 256, 512], nb=2, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") model.load_state_dict(torch.load(model_path), strict=True) model.eval() for key, v in model.named_parameters(): v.requires_grad = False number_parameters = sum(map(lambda x: x.numel(), model.parameters())) logger.info('Params number: {}'.format(number_parameters)) model = model.to(device) logger.info('Model path: {:s}'.format(model_path)) logger.info('model_name:{}, image sigma:{}'.format(model_name, noise_level_img)) logger.info(L_path) img = os.path.join(L_path, test_image) # ------------------------------------ # (1) img_L # ------------------------------------ img_name, ext = os.path.splitext(os.path.basename(img)) img_L = util.imread_uint(img, n_channels=n_channels) img_L = util.uint2single(img_L) util.imshow(img_L) if show_img else None w, h = img_L.shape[:2] logger.info('{:>10s}--> ({:>4d}x{:<4d})'.format(img_name+ext, w, h)) # boundary handling boarder = 8 # default setting for kernel size 25x25 img = cv2.resize(img_L, (sf*h, sf*w), interpolation=cv2.INTER_NEAREST) img = utils_deblur.wrap_boundary_liu(img, [int(np.ceil(sf*w/boarder+2)*boarder), int(np.ceil(sf*h/boarder+2)*boarder)]) img_wrap = sr.downsample_np(img, sf, center=False) img_wrap[:w, :h, :] = img_L img_L = img_wrap util.imshow(util.single2uint(img_L), title='LR image with noise level {}'.format(noise_level_img)) if show_img else None img_L = util.single2tensor4(img_L) img_L = img_L.to(device) # ------------------------------------ # (2) img_E # ------------------------------------ sigma = torch.tensor(noise_level_model).float().view([1, 1, 1, 1]) [img_L, kernel, sigma] = [el.to(device) for el in [img_L, kernel, sigma]] img_E = model(img_L, kernel, sf, sigma) img_E = util.tensor2uint(img_E)[:sf*w, :sf*h, ...] if save_E: util.imsave(img_E, os.path.join(E_path, img_name+'_x'+str(sf)+'_'+model_name+'.png')) # -------------------------------- # (3) save img_LE # -------------------------------- if save_LE: k_v = k/np.max(k)*1.2 k_v = util.single2uint(np.tile(k_v[..., np.newaxis], [1, 1, 3])) k_factor = 3 k_v = cv2.resize(k_v, (k_factor*k_v.shape[1], k_factor*k_v.shape[0]), interpolation=cv2.INTER_NEAREST) img_L = util.tensor2uint(img_L)[:w, :h, ...] img_I = cv2.resize(img_L, (sf*img_L.shape[1], sf*img_L.shape[0]), interpolation=cv2.INTER_NEAREST) img_I[:k_v.shape[0], :k_v.shape[1], :] = k_v util.imshow(np.concatenate([img_I, img_E], axis=1), title='LR / Recovered') if show_img else None util.imsave(np.concatenate([img_I, img_E], axis=1), os.path.join(E_path, img_name+'_x'+str(sf)+'_'+model_name+'_LE.png'))
def __getitem__(self, index): # ------------------------------------ # get H image # ------------------------------------ H_path = self.paths_H[index] img_H = util.imread_uint(H_path, self.n_channels) L_path = H_path if self.opt['phase'] == 'train': """ # -------------------------------- # get L/H patch pairs # -------------------------------- """ H, W, _ = img_H.shape # -------------------------------- # randomly crop the patch # -------------------------------- rnd_h = random.randint(0, max(0, H - self.patch_size)) rnd_w = random.randint(0, max(0, W - self.patch_size)) patch_H = img_H[rnd_h:rnd_h + self.patch_size, rnd_w:rnd_w + self.patch_size, :] # -------------------------------- # augmentation - flip, rotate # -------------------------------- mode = np.random.randint(0, 8) patch_H = util.augment_img(patch_H, mode=mode) # -------------------------------- # HWC to CHW, numpy(uint) to tensor # -------------------------------- img_H = util.uint2tensor3(patch_H) img_L = img_H.clone() # -------------------------------- # add noise # -------------------------------- noise = torch.randn(img_L.size()).mul_(self.sigma / 255.0) img_L.add_(noise) else: """ # -------------------------------- # get L/H image pairs # -------------------------------- """ img_H = util.uint2single(img_H) img_L = np.copy(img_H) # -------------------------------- # add noise # -------------------------------- np.random.seed(seed=0) img_L += np.random.normal(0, self.sigma_test / 255.0, img_L.shape) # -------------------------------- # HWC to CHW, numpy to tensor # -------------------------------- img_L = util.single2tensor3(img_L) img_H = util.single2tensor3(img_H) return {'L': img_L, 'H': img_H, 'H_path': H_path, 'L_path': L_path}
def main(): # ---------------------------------------- # Preparation # ---------------------------------------- model_name = 'usrnet' # 'usrgan' | 'usrnet' | 'usrgan_tiny' | 'usrnet_tiny' testset_name = 'set5' # test set, 'set5' | 'srbsd68' need_degradation = True # default: True sf = 4 # scale factor, only from {2, 3, 4} show_img = False # default: False save_L = True # save LR image save_E = True # save estimated image # load approximated bicubic kernels #kernels = hdf5storage.loadmat(os.path.join('kernels', 'kernels_bicubicx234.mat'))['kernels'] kernels = loadmat(os.path.join('kernels', 'kernels_bicubicx234.mat'))['kernels'] kernel = kernels[0, sf - 2].astype(np.float64) kernel = util.single2tensor4(kernel[..., np.newaxis]) task_current = 'sr' # fixed, 'sr' for super-resolution n_channels = 3 # fixed, 3 for color image model_pool = 'model_zoo' # fixed testsets = 'testsets' # fixed results = 'results' # fixed noise_level_img = 0 # fixed: 0, noise level for LR image noise_level_model = noise_level_img # fixed, noise level of model, default 0 result_name = testset_name + '_' + model_name + '_bicubic' border = sf if task_current == 'sr' else 0 # shave boader to calculate PSNR and SSIM model_path = os.path.join(model_pool, model_name + '.pth') # ---------------------------------------- # L_path, E_path, H_path # ---------------------------------------- L_path = os.path.join( testsets, testset_name) # L_path, fixed, for Low-quality images H_path = L_path # H_path, 'None' | L_path, for High-quality images E_path = os.path.join(results, result_name) # E_path, fixed, for Estimated images util.mkdir(E_path) if H_path == L_path: need_degradation = True logger_name = result_name utils_logger.logger_info(logger_name, log_path=os.path.join(E_path, logger_name + '.log')) logger = logging.getLogger(logger_name) need_H = True if H_path is not None else False device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') # ---------------------------------------- # load model # ---------------------------------------- from models.network_usrnet import USRNet as net # for pytorch version <= 1.7.1 # from models.network_usrnet_v1 import USRNet as net # for pytorch version >=1.8.1 if 'tiny' in model_name: model = net(n_iter=6, h_nc=32, in_nc=4, out_nc=3, nc=[16, 32, 64, 64], nb=2, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") else: model = net(n_iter=8, h_nc=64, in_nc=4, out_nc=3, nc=[64, 128, 256, 512], nb=2, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") model.load_state_dict(torch.load(model_path), strict=True) model.eval() for key, v in model.named_parameters(): v.requires_grad = False number_parameters = sum(map(lambda x: x.numel(), model.parameters())) logger.info('Params number: {}'.format(number_parameters)) model = model.to(device) logger.info('Model path: {:s}'.format(model_path)) test_results = OrderedDict() test_results['psnr'] = [] test_results['ssim'] = [] test_results['psnr_y'] = [] test_results['ssim_y'] = [] logger.info('model_name:{}, image sigma:{}'.format(model_name, noise_level_img)) logger.info(L_path) L_paths = util.get_image_paths(L_path) H_paths = util.get_image_paths(H_path) if need_H else None for idx, img in enumerate(L_paths): # ------------------------------------ # (1) img_L # ------------------------------------ img_name, ext = os.path.splitext(os.path.basename(img)) logger.info('{:->4d}--> {:>10s}'.format(idx + 1, img_name + ext)) img_L = util.imread_uint(img, n_channels=n_channels) img_L = util.uint2single(img_L) # degradation process, bicubic downsampling if need_degradation: img_L = util.modcrop(img_L, sf) img_L = util.imresize_np(img_L, 1 / sf) # img_L = util.uint2single(util.single2uint(img_L)) # np.random.seed(seed=0) # for reproducibility # img_L += np.random.normal(0, noise_level_img/255., img_L.shape) w, h = img_L.shape[:2] if save_L: util.imsave( util.single2uint(img_L), os.path.join(E_path, img_name + '_LR_x' + str(sf) + '.png')) img = cv2.resize(img_L, (sf * h, sf * w), interpolation=cv2.INTER_NEAREST) img = utils_deblur.wrap_boundary_liu(img, [ int(np.ceil(sf * w / 8 + 2) * 8), int(np.ceil(sf * h / 8 + 2) * 8) ]) img_wrap = sr.downsample_np(img, sf, center=False) img_wrap[:w, :h, :] = img_L img_L = img_wrap util.imshow(util.single2uint(img_L), title='LR image with noise level {}'.format( noise_level_img)) if show_img else None img_L = util.single2tensor4(img_L) img_L = img_L.to(device) # ------------------------------------ # (2) img_E # ------------------------------------ sigma = torch.tensor(noise_level_model).float().view([1, 1, 1, 1]) [img_L, kernel, sigma] = [el.to(device) for el in [img_L, kernel, sigma]] img_E = model(img_L, kernel, sf, sigma) img_E = util.tensor2uint(img_E) img_E = img_E[:sf * w, :sf * h, :] if need_H: # -------------------------------- # (3) img_H # -------------------------------- img_H = util.imread_uint(H_paths[idx], n_channels=n_channels) img_H = img_H.squeeze() img_H = util.modcrop(img_H, sf) # -------------------------------- # PSNR and SSIM # -------------------------------- psnr = util.calculate_psnr(img_E, img_H, border=border) ssim = util.calculate_ssim(img_E, img_H, border=border) test_results['psnr'].append(psnr) test_results['ssim'].append(ssim) logger.info('{:s} - PSNR: {:.2f} dB; SSIM: {:.4f}.'.format( img_name + ext, psnr, ssim)) util.imshow(np.concatenate([img_E, img_H], axis=1), title='Recovered / Ground-truth') if show_img else None if np.ndim(img_H) == 3: # RGB image img_E_y = util.rgb2ycbcr(img_E, only_y=True) img_H_y = util.rgb2ycbcr(img_H, only_y=True) psnr_y = util.calculate_psnr(img_E_y, img_H_y, border=border) ssim_y = util.calculate_ssim(img_E_y, img_H_y, border=border) test_results['psnr_y'].append(psnr_y) test_results['ssim_y'].append(ssim_y) # ------------------------------------ # save results # ------------------------------------ if save_E: util.imsave( img_E, os.path.join( E_path, img_name + '_x' + str(sf) + '_' + model_name + '.png')) if need_H: ave_psnr = sum(test_results['psnr']) / len(test_results['psnr']) ave_ssim = sum(test_results['ssim']) / len(test_results['ssim']) logger.info( 'Average PSNR/SSIM(RGB) - {} - x{} --PSNR: {:.2f} dB; SSIM: {:.4f}' .format(result_name, sf, ave_psnr, ave_ssim)) if np.ndim(img_H) == 3: ave_psnr_y = sum(test_results['psnr_y']) / len( test_results['psnr_y']) ave_ssim_y = sum(test_results['ssim_y']) / len( test_results['ssim_y']) logger.info( 'Average PSNR/SSIM( Y ) - {} - x{} - PSNR: {:.2f} dB; SSIM: {:.4f}' .format(result_name, sf, ave_psnr_y, ave_ssim_y))
epochs = 50 model_path = os.path.join(model_pool, model_name+'.pth') test_path = os.path.join(test_im) train_path = os.path.join(train_im) # load model model = DnCNN(in_nc=n_channels, out_nc=n_channels, nc=64, nb=17, act_mode='R') model.load_state_dict(torch.load(model_path), strict=True) model = model.to(device) model.eval() # load test image x = util.imread_uint(test_path, n_channels=n_channels) orig_im = x.squeeze() x = util.uint2single(x) np.random.seed(seed=0) # for reproducibility y = x + np.random.normal(0, sigma/255., x.shape) # add gaussian noise y = util.single2tensor4(y) y = y.to(device) # denoise the image to compare PSNR before and after adaptation with torch.no_grad(): x_ = model(y) # compute PSNR denoised_im = util.tensor2uint(x_) prev_psnr = util.calculate_psnr(denoised_im, orig_im, border=0) # external adaptation
def main(): # ---------------------------------------- # Preparation # ---------------------------------------- model_name = 'dncnn3' # 'dncnn3'- can be used for blind Gaussian denoising, JPEG deblocking (quality factor 5-100) and super-resolution (x234) # important! testset_name = 'bsd68' # test set, low-quality grayscale/color JPEG images n_channels = 1 # set 1 for grayscale image, set 3 for color image x8 = False # default: False, x8 to boost performance testsets = 'testsets' # fixed results = 'results' # fixed result_name = testset_name + '_' + model_name # fixed L_path = os.path.join( testsets, testset_name ) # L_path, for Low-quality grayscale/Y-channel JPEG images E_path = os.path.join(results, result_name) # E_path, for Estimated images util.mkdir(E_path) model_pool = 'model_zoo' # fixed model_path = os.path.join(model_pool, model_name + '.pth') logger_name = result_name utils_logger.logger_info(logger_name, log_path=os.path.join(E_path, logger_name + '.log')) logger = logging.getLogger(logger_name) device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') # ---------------------------------------- # load model # ---------------------------------------- from models.network_dncnn import DnCNN as net model = net(in_nc=1, out_nc=1, nc=64, nb=20, act_mode='R') model.load_state_dict(torch.load(model_path), strict=True) model.eval() for k, v in model.named_parameters(): v.requires_grad = False model = model.to(device) logger.info('Model path: {:s}'.format(model_path)) number_parameters = sum(map(lambda x: x.numel(), model.parameters())) logger.info('Params number: {}'.format(number_parameters)) logger.info(L_path) L_paths = util.get_image_paths(L_path) for idx, img in enumerate(L_paths): # ------------------------------------ # (1) img_L # ------------------------------------ img_name, ext = os.path.splitext(os.path.basename(img)) logger.info('{:->4d}--> {:>10s}'.format(idx + 1, img_name + ext)) img_L = util.imread_uint(img, n_channels=n_channels) img_L = util.uint2single(img_L) if n_channels == 3: ycbcr = util.rgb2ycbcr(img_L, False) img_L = ycbcr[..., 0:1] img_L = util.single2tensor4(img_L) img_L = img_L.to(device) # ------------------------------------ # (2) img_E # ------------------------------------ if not x8: img_E = model(img_L) else: img_E = utils_model.test_mode(model, img_L, mode=3) img_E = util.tensor2single(img_E) if n_channels == 3: ycbcr[..., 0] = img_E img_E = util.ycbcr2rgb(ycbcr) img_E = util.single2uint(img_E) # ------------------------------------ # save results # ------------------------------------ util.imsave(img_E, os.path.join(E_path, img_name + '.png'))
def main(): # ---------------------------------------- # load kernels # ---------------------------------------- #PSF_grid = np.load('./data/AC254-075-A-ML-Zemax(ZMX).npz')['PSF'] PSF_grid = np.load('./data/Heide_PSF_plano_small.npz')['PSF'] PSF_grid = PSF_grid.astype(np.float32) gx, gy = PSF_grid.shape[:2] for xx in range(gx): for yy in range(gy): PSF_grid[xx, yy] = PSF_grid[xx, yy] / np.sum(PSF_grid[xx, yy], axis=(0, 1)) # ---------------------------------------- # load model # ---------------------------------------- stage = 8 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') model = net(n_iter=stage, h_nc=64, in_nc=4, out_nc=3, nc=[64, 128, 256, 512], nb=2, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") model_code = 'iter800' loaded_state = torch.load( '/home/xiu/databag/deblur/models/plano/uabcnet_{}.pth'.format( model_code)) #strip_state = strip_prefix_if_present(loaded_state,prefix="p.") model.load_state_dict(loaded_state, strict=True) model.eval() for _, v in model.named_parameters(): v.requires_grad = False model = model.to(device) for img_id in range(1, 237): #for img_id in range(1,12): #img_L = cv2.imread('/home/xiu/workspace/UABC/ICCV2021/video1-3/res/2_{:03d}.bmp'.format(img_id)) #img_L = cv2.imread('/home/xiu/workspace/UABC/ICCV2021/video/{:08d}.bmp'.format(img_id)) #img_L = cv2.imread('/home/xiu/databag/deblur/ICCV2021/suo_image/{}/AC254-075-A-ML-Zemax(ZMX).bmp'.format(img_id)) #img_L = cv2.imread('/home/xiu/workspace/UABC/ICCV2021/ResolutionChart/Reso.bmp') img_L = cv2.imread( '/home/xiu/databag/deblur/ICCV2021/MPI_data/drain/blurry.jpg') img_L = img_L.astype(np.float32) img_L = img_L[38:-39, 74:-74] img_L = cv2.resize(img_L, dsize=None, fx=0.5, fy=0.5) #img_L = np.pad(img_L,((1,1),(61,62),(0,0)),mode='edge') W, H = img_L.shape[:2] print(gx, gy) num_patch = [gx, gy] #positional alpha-beta parameters for HQS ab_numpy = np.loadtxt( '/home/xiu/databag/deblur/models/plano/ab_{}.txt'.format( model_code)).astype(np.float32).reshape(gx, gy, stage * 2, 3) ab = torch.tensor(ab_numpy, device=device, requires_grad=False) t0 = time.time() px_start = 0 py_start = 0 PSF_patch = PSF_grid[px_start:px_start + num_patch[0], py_start:py_start + num_patch[1]] #block_expand = 1 patch_L = img_L[px_start * W // gx:(px_start + num_patch[0]) * W // gx, py_start * H // gy:(py_start + num_patch[1]) * H // gy, :] p_W, p_H = patch_L.shape[:2] expand = max(PSF_grid.shape[2] // 2, p_W // 16) block_expand = expand patch_L_wrap = util_deblur.wrap_boundary_liu( patch_L, (p_W + block_expand * 2, p_H + block_expand * 2)) #centralize patch_L_wrap = np.hstack((patch_L_wrap[:, -block_expand:, :], patch_L_wrap[:, :p_H + block_expand, :])) patch_L_wrap = np.vstack((patch_L_wrap[-block_expand:, :, :], patch_L_wrap[:p_W + block_expand, :, :])) x = util.uint2single(patch_L_wrap) x = util.single2tensor4(x) k_all = [] for h_ in range(num_patch[1]): for w_ in range(num_patch[0]): k_all.append(util.single2tensor4(PSF_patch[w_, h_])) k = torch.cat(k_all, dim=0) [x, k] = [el.to(device) for el in [x, k]] ab_patch = F.softplus(ab[px_start:px_start + num_patch[0], py_start:py_start + num_patch[1]]) cd = [] for h_ in range(num_patch[1]): for w_ in range(num_patch[0]): cd.append(ab_patch[w_:w_ + 1, h_]) cd = torch.cat(cd, dim=0) x_E = model.forward_patchwise(x, k, cd, num_patch, [W // gx, H // gy]) x_E = x_E[..., block_expand:block_expand + p_W, block_expand:block_expand + p_H] patch_L = patch_L_wrap.astype(np.uint8) patch_E = util.tensor2uint(x_E) #patch_E_z = np.hstack((patch_E_all[::2])) #patch_E_x = np.hstack((patch_E_all[1::2])) #patch_E_show = np.vstack((patch_E_z,patch_E_x)) #if block_expand>0: # show = np.hstack((patch_L[block_expand:-block_expand,block_expand:-block_expand],patch_E)) #else: # show = np.hstack((patch_L,patch_E)) #cv2.imshow('stage',patch_E_show) #cv2.imshow('HL',show) #cv2.imshow('RGB',rgb) #key = cv2.waitKey(-1) #if key==ord('n'): # break t1 = time.time() print(t1 - t0) # print(i) xk = patch_E # #zk = zk.astype(np.uint8) xk = xk.astype(np.uint8) #cv2.imwrite('/home/xiu/workspace/UABC/ICCV2021/new_image/image/ours-{}.png'.format(img_id),xk) #cv2.imwrite('/home/xiu/workspace/UABC/ICCV2021/video_deblur/{:08d}.png'.format(img_id),xk) #cv2.imwrite('/home/xiu/workspace/UABC/ICCV2021/cap_result/1_{:03d}.png'.format(img_id),xk) cv2.imshow('xx', xk) cv2.imshow('img_L', patch_L.astype(np.uint8)) key = cv2.waitKey(-1) if key == ord('q'): break
def main(): # ---------------------------------------- # Preparation # ---------------------------------------- noise_level_img = 15 # set AWGN noise level for noisy image noise_level_model = noise_level_img # set noise level for model model_name = 'drunet_gray' # set denoiser model, 'drunet_gray' | 'drunet_color' testset_name = 'bsd68' # set test set, 'bsd68' | 'cbsd68' | 'set12' x8 = False # default: False, x8 to boost performance show_img = False # default: False border = 0 # shave boader to calculate PSNR and SSIM if 'color' in model_name: n_channels = 3 # 3 for color image else: n_channels = 1 # 1 for grayscale image model_pool = 'model_zoo' # fixed testsets = 'testsets' # fixed results = 'results' # fixed task_current = 'dn' # 'dn' for denoising result_name = testset_name + '_' + task_current + '_' + model_name model_path = os.path.join(model_pool, model_name + '.pth') device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') torch.cuda.empty_cache() # ---------------------------------------- # L_path, E_path, H_path # ---------------------------------------- L_path = os.path.join(testsets, testset_name) # L_path, for Low-quality images E_path = os.path.join(results, result_name) # E_path, for Estimated images util.mkdir(E_path) logger_name = result_name utils_logger.logger_info(logger_name, log_path=os.path.join(E_path, logger_name + '.log')) logger = logging.getLogger(logger_name) # ---------------------------------------- # load model # ---------------------------------------- from models.network_unet import UNetRes as net model = net(in_nc=n_channels + 1, out_nc=n_channels, nc=[64, 128, 256, 512], nb=4, act_mode='R', downsample_mode="strideconv", upsample_mode="convtranspose") model.load_state_dict(torch.load(model_path), strict=True) model.eval() for k, v in model.named_parameters(): v.requires_grad = False model = model.to(device) logger.info('Model path: {:s}'.format(model_path)) number_parameters = sum(map(lambda x: x.numel(), model.parameters())) logger.info('Params number: {}'.format(number_parameters)) test_results = OrderedDict() test_results['psnr'] = [] test_results['ssim'] = [] logger.info('model_name:{}, model sigma:{}, image sigma:{}'.format( model_name, noise_level_img, noise_level_model)) logger.info(L_path) L_paths = util.get_image_paths(L_path) for idx, img in enumerate(L_paths): # ------------------------------------ # (1) img_L # ------------------------------------ img_name, ext = os.path.splitext(os.path.basename(img)) # logger.info('{:->4d}--> {:>10s}'.format(idx+1, img_name+ext)) img_H = util.imread_uint(img, n_channels=n_channels) img_L = util.uint2single(img_H) # Add noise without clipping np.random.seed(seed=0) # for reproducibility img_L += np.random.normal(0, noise_level_img / 255., img_L.shape) util.imshow(util.single2uint(img_L), title='Noisy image with noise level {}'.format( noise_level_img)) if show_img else None img_L = util.single2tensor4(img_L) img_L = torch.cat( (img_L, torch.FloatTensor([noise_level_model / 255.]).repeat( 1, 1, img_L.shape[2], img_L.shape[3])), dim=1) img_L = img_L.to(device) # ------------------------------------ # (2) img_E # ------------------------------------ if not x8 and img_L.size(2) // 8 == 0 and img_L.size(3) // 8 == 0: img_E = model(img_L) elif not x8 and (img_L.size(2) // 8 != 0 or img_L.size(3) // 8 != 0): img_E = utils_model.test_mode(model, img_L, refield=64, mode=5) elif x8: img_E = utils_model.test_mode(model, img_L, mode=3) img_E = util.tensor2uint(img_E) # -------------------------------- # PSNR and SSIM # -------------------------------- if n_channels == 1: img_H = img_H.squeeze() psnr = util.calculate_psnr(img_E, img_H, border=border) ssim = util.calculate_ssim(img_E, img_H, border=border) test_results['psnr'].append(psnr) test_results['ssim'].append(ssim) logger.info('{:s} - PSNR: {:.2f} dB; SSIM: {:.4f}.'.format( img_name + ext, psnr, ssim)) # ------------------------------------ # save results # ------------------------------------ util.imsave(img_E, os.path.join(E_path, img_name + ext)) ave_psnr = sum(test_results['psnr']) / len(test_results['psnr']) ave_ssim = sum(test_results['ssim']) / len(test_results['ssim']) logger.info( 'Average PSNR/SSIM(RGB) - {} - PSNR: {:.2f} dB; SSIM: {:.4f}'.format( result_name, ave_psnr, ave_ssim))
def main(): # ---------------------------------------- # load kernels # ---------------------------------------- #PSF_grid = np.load('./data/Schuler_PSF01.npz')['PSF'] #PSF_grid = np.load('./data/Schuler_PSF_facade.npz')['PSF'] PSF_grid = np.load('./data/ZEMAX-AC254-075-A-new.npz')['PSF'] #PSF_grid = np.load('./data/Schuler_PSF03.npz')['PSF'] #PSF_grid = np.load('./data/PSF.npz')['PSF'] #print(PSF_grid.shape) PSF_grid = PSF_grid.astype(np.float32) gx,gy = PSF_grid.shape[:2] for xx in range(gx): for yy in range(gy): PSF_grid[xx,yy] = PSF_grid[xx,yy]/np.sum(PSF_grid[xx,yy],axis=(0,1)) #PSF_grid = PSF_grid[:,1:-1,...] # ---------------------------------------- # load model # ---------------------------------------- device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') model = net(n_iter=8, h_nc=64, in_nc=4, out_nc=3, nc=[64, 128, 256, 512], nb=2, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") loaded_state = torch.load('./usrnet_ZEMAX.pth') #strip_state = strip_prefix_if_present(loaded_state,prefix="p.") model.load_state_dict(loaded_state, strict=True) model.eval() #model.train() for _, v in model.named_parameters(): v.requires_grad = False # v.requires_grad = False model = model.to(device) for img_id in range(100): img_L = cv2.imread('/home/xiu/workspace/UABC/ICCV2021/video/{:08d}.bmp'.format(img_id)) img_L = img_L.astype(np.float32) img_E = np.zeros_like(img_L) weight_E = np.zeros_like(img_L) patch_size = 2*128 num_patch = 2 p_size = patch_size//num_patch expand = PSF_grid.shape[2] ab_numpy = np.loadtxt('ab_ZEMAX.txt').astype(np.float32).reshape(6,8,16,3) ab = torch.tensor(ab_numpy,device=device,requires_grad=False) #save img_L t0 = time.time() #while running: for px_start in range(0,6-2+1,2): for py_start in range(0,8-2+1,2): PSF_patch = PSF_grid[px_start:px_start+num_patch,py_start:py_start+num_patch] patch_L = img_L[px_start*p_size:(px_start+num_patch)*p_size,py_start*p_size:py_start*p_size+num_patch*p_size,:] block_expand = expand #block_expand = 1 if block_expand > 0: patch_L_wrap = util_deblur.wrap_boundary_liu(patch_L,(patch_size+block_expand*2,patch_size+block_expand*2)) #centralize patch_L_wrap = np.hstack((patch_L_wrap[:,-block_expand:,:],patch_L_wrap[:,:patch_size+block_expand,:])) patch_L_wrap = np.vstack((patch_L_wrap[-block_expand:,:,:],patch_L_wrap[:patch_size+block_expand,:,:])) else: patch_L_wrap = patch_L if block_expand>0: x = util.uint2single(patch_L_wrap) else: x = util.uint2single(patch_L) x = util.single2tensor4(x) # x_blocky = torch.cat(torch.chunk(x,num_patch,dim=2),dim=0) # x_blocky = torch.cat(torch.chunk(x_blocky,num_patch,dim=3),dim=0) k_all = [] for w_ in range(num_patch): for h_ in range(num_patch): k_all.append(util.single2tensor4(PSF_patch[h_,w_])) k = torch.cat(k_all,dim=0) [x,k] = [el.to(device) for el in [x,k]] cd = F.softplus(ab[px_start:px_start+num_patch,py_start:py_start+num_patch]) cd = cd.view(num_patch**2,2*8,3) x_E = model.forward_patchwise(x,k,cd,[num_patch,num_patch],[patch_size//num_patch,patch_size//num_patch]) patch_L = patch_L_wrap.astype(np.uint8) patch_E = util.tensor2uint(x_E) patch_E_all = [util.tensor2uint(pp) for pp in x_E] #patch_E_z = np.hstack((patch_E_all[::2])) #patch_E_x = np.hstack((patch_E_all[1::2])) #patch_E_show = np.vstack((patch_E_z,patch_E_x)) #if block_expand>0: # show = np.hstack((patch_L[block_expand:-block_expand,block_expand:-block_expand],patch_E)) #else: # show = np.hstack((patch_L,patch_E)) #get kernel for i in range(8): img_E_deconv[i][px_start*p_size:(px_start+num_patch)*p_size,py_start*p_size:py_start*p_size+num_patch*p_size,:] += patch_E_all[-2][expand:-expand,expand:-expand] img_E_denoise[i][px_start*p_size:(px_start+num_patch)*p_size,py_start*p_size:py_start*p_size+num_patch*p_size,:] += patch_E_all[-1][expand:-expand,expand:-expand] weight_E[px_start*p_size:(px_start+num_patch)*p_size,py_start*p_size:py_start*p_size+num_patch*p_size,:] += 1.0 #cv2.imshow('stage',patch_E_show) #cv2.imshow('HL',show) #cv2.imshow('RGB',rgb) #key = cv2.waitKey(-1) #if key==ord('n'): # break t1 = time.time() print(t1-t0) img_E = img_E/weight_E img_E_deconv = [pp/weight_E for pp in img_E_deconv] img_E_denoise = [pp/weight_E for pp in img_E_denoise] # print(i) xk = img_E_denoise[-1] # #zk = zk.astype(np.uint8) xk = xk.astype(np.uint8) #cv2.imwrite('/home/xiu/workspace/UABC/ICCV2021/video_deblur/{:08d}.png'.format(img_id),xk) cv2.imshow('xx',xk) cv2.imshow('img_L',img_L.astype(np.uint8)) cv2.waitKey(-1)
def main(): # ---------------------------------------- # Preparation # ---------------------------------------- noise_level_img = 7.65 / 255.0 # default: 0, noise level for LR image noise_level_model = noise_level_img # noise level of model, default 0 model_name = 'drunet_gray' # 'drunet_gray' | 'drunet_color' | 'ircnn_gray' | 'ircnn_color' testset_name = 'Set3C' # test set, 'set5' | 'srbsd68' x8 = True # default: False, x8 to boost performance iter_num = 8 # number of iterations modelSigma1 = 49 modelSigma2 = noise_level_model * 255. show_img = False # default: False save_L = True # save LR image save_E = True # save estimated image save_LEH = False # save zoomed LR, E and H images border = 0 # -------------------------------- # load kernel # -------------------------------- kernels = hdf5storage.loadmat(os.path.join('kernels', 'Levin09.mat'))['kernels'] sf = 1 task_current = 'deblur' # 'deblur' for deblurring n_channels = 3 if 'color' in model_name else 1 # fixed model_zoo = 'model_zoo' # fixed testsets = 'testsets' # fixed results = 'results' # fixed result_name = testset_name + '_' + task_current + '_' + model_name model_path = os.path.join(model_zoo, model_name + '.pth') device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') torch.cuda.empty_cache() # ---------------------------------------- # L_path, E_path, H_path # ---------------------------------------- L_path = os.path.join(testsets, testset_name) # L_path, for Low-quality images E_path = os.path.join(results, result_name) # E_path, for Estimated images util.mkdir(E_path) logger_name = result_name utils_logger.logger_info(logger_name, log_path=os.path.join(E_path, logger_name + '.log')) logger = logging.getLogger(logger_name) # ---------------------------------------- # load model # ---------------------------------------- if 'drunet' in model_name: from models.network_unet import UNetRes as net model = net(in_nc=n_channels + 1, out_nc=n_channels, nc=[64, 128, 256, 512], nb=4, act_mode='R', downsample_mode="strideconv", upsample_mode="convtranspose") model.load_state_dict(torch.load(model_path), strict=True) model.eval() for _, v in model.named_parameters(): v.requires_grad = False model = model.to(device) elif 'ircnn' in model_name: from models.network_dncnn import IRCNN as net model = net(in_nc=n_channels, out_nc=n_channels, nc=64) model25 = torch.load(model_path) former_idx = 0 logger.info('model_name:{}, image sigma:{:.3f}, model sigma:{:.3f}'.format( model_name, noise_level_img, noise_level_model)) logger.info('Model path: {:s}'.format(model_path)) logger.info(L_path) L_paths = util.get_image_paths(L_path) test_results_ave = OrderedDict() test_results_ave['psnr'] = [] # record average PSNR for each kernel for k_index in range(kernels.shape[1]): logger.info('-------k:{:>2d} ---------'.format(k_index)) test_results = OrderedDict() test_results['psnr'] = [] k = kernels[0, k_index].astype(np.float64) util.imshow(k) if show_img else None for idx, img in enumerate(L_paths): # -------------------------------- # (1) get img_L # -------------------------------- img_name, ext = os.path.splitext(os.path.basename(img)) img_H = util.imread_uint(img, n_channels=n_channels) img_H = util.modcrop(img_H, 8) # modcrop img_L = ndimage.filters.convolve(img_H, np.expand_dims(k, axis=2), mode='wrap') util.imshow(img_L) if show_img else None img_L = util.uint2single(img_L) np.random.seed(seed=0) # for reproducibility img_L += np.random.normal(0, noise_level_img, img_L.shape) # add AWGN # -------------------------------- # (2) get rhos and sigmas # -------------------------------- rhos, sigmas = pnp.get_rho_sigma(sigma=max(0.255 / 255., noise_level_model), iter_num=iter_num, modelSigma1=modelSigma1, modelSigma2=modelSigma2, w=1.0) rhos, sigmas = torch.tensor(rhos).to(device), torch.tensor( sigmas).to(device) # -------------------------------- # (3) initialize x, and pre-calculation # -------------------------------- x = util.single2tensor4(img_L).to(device) img_L_tensor, k_tensor = util.single2tensor4( img_L), util.single2tensor4(np.expand_dims(k, 2)) [k_tensor, img_L_tensor] = util.todevice([k_tensor, img_L_tensor], device) FB, FBC, F2B, FBFy = sr.pre_calculate(img_L_tensor, k_tensor, sf) # -------------------------------- # (4) main iterations # -------------------------------- for i in range(iter_num): # -------------------------------- # step 1, FFT # -------------------------------- tau = rhos[i].float().repeat(1, 1, 1, 1) x = sr.data_solution(x, FB, FBC, F2B, FBFy, tau, sf) if 'ircnn' in model_name: current_idx = np.int( np.ceil(sigmas[i].cpu().numpy() * 255. / 2.) - 1) if current_idx != former_idx: model.load_state_dict(model25[str(current_idx)], strict=True) model.eval() for _, v in model.named_parameters(): v.requires_grad = False model = model.to(device) former_idx = current_idx # -------------------------------- # step 2, denoiser # -------------------------------- if x8: x = util.augment_img_tensor4(x, i % 8) if 'drunet' in model_name: x = torch.cat((x, sigmas[i].float().repeat( 1, 1, x.shape[2], x.shape[3])), dim=1) x = utils_model.test_mode(model, x, mode=2, refield=32, min_size=256, modulo=16) elif 'ircnn' in model_name: x = model(x) if x8: if i % 8 == 3 or i % 8 == 5: x = util.augment_img_tensor4(x, 8 - i % 8) else: x = util.augment_img_tensor4(x, i % 8) # -------------------------------- # (3) img_E # -------------------------------- img_E = util.tensor2uint(x) if n_channels == 1: img_H = img_H.squeeze() if save_E: util.imsave( img_E, os.path.join( E_path, img_name + '_k' + str(k_index) + '_' + model_name + '.png')) # -------------------------------- # (4) img_LEH # -------------------------------- if save_LEH: img_L = util.single2uint(img_L) k_v = k / np.max(k) * 1.0 k_v = util.single2uint(np.tile(k_v[..., np.newaxis], [1, 1, 3])) k_v = cv2.resize(k_v, (3 * k_v.shape[1], 3 * k_v.shape[0]), interpolation=cv2.INTER_NEAREST) img_I = cv2.resize(img_L, (sf * img_L.shape[1], sf * img_L.shape[0]), interpolation=cv2.INTER_NEAREST) img_I[:k_v.shape[0], -k_v.shape[1]:, :] = k_v img_I[:img_L.shape[0], :img_L.shape[1], :] = img_L util.imshow(np.concatenate([img_I, img_E, img_H], axis=1), title='LR / Recovered / Ground-truth' ) if show_img else None util.imsave( np.concatenate([img_I, img_E, img_H], axis=1), os.path.join(E_path, img_name + '_k' + str(k_index) + '_LEH.png')) if save_L: util.imsave( util.single2uint(img_L), os.path.join(E_path, img_name + '_k' + str(k_index) + '_LR.png')) psnr = util.calculate_psnr( img_E, img_H, border=border) # change with your own border test_results['psnr'].append(psnr) logger.info('{:->4d}--> {:>10s} --k:{:>2d} PSNR: {:.2f}dB'.format( idx + 1, img_name + ext, k_index, psnr)) # -------------------------------- # Average PSNR # -------------------------------- ave_psnr = sum(test_results['psnr']) / len(test_results['psnr']) logger.info( '------> Average PSNR of ({}), kernel: ({}) sigma: ({:.2f}): {:.2f} dB' .format(testset_name, k_index, noise_level_model, ave_psnr)) test_results_ave['psnr'].append(ave_psnr)
def main(): # ---------------------------------------- # Preparation # ---------------------------------------- noise_level_img = 0 # default: 0, noise level for LR image noise_level_model = noise_level_img # noise level for model model_name = 'dpsr_x4_gan' # 'dpsr_x2' | 'dpsr_x3' | 'dpsr_x4' | 'dpsr_x4_gan' testset_name = 'set5' # test set, 'set5' | 'srbsd68' need_degradation = True # default: True x8 = False # default: False, x8 to boost performance sf = [int(s) for s in re.findall(r'\d+', model_name)][0] # scale factor show_img = False # default: False task_current = 'sr' # 'dn' for denoising | 'sr' for super-resolution n_channels = 3 # fixed nc = 96 # fixed, number of channels nb = 16 # fixed, number of conv layers model_pool = 'model_zoo' # fixed testsets = 'testsets' # fixed results = 'results' # fixed result_name = testset_name + '_' + model_name border = sf if task_current == 'sr' else 0 # shave boader to calculate PSNR and SSIM model_path = os.path.join(model_pool, model_name + '.pth') # ---------------------------------------- # L_path, E_path, H_path # ---------------------------------------- L_path = os.path.join(testsets, testset_name) # L_path, for Low-quality images H_path = L_path # H_path, for High-quality images E_path = os.path.join(results, result_name) # E_path, for Estimated images util.mkdir(E_path) if H_path == L_path: need_degradation = True logger_name = result_name utils_logger.logger_info(logger_name, log_path=os.path.join(E_path, logger_name + '.log')) logger = logging.getLogger(logger_name) need_H = True if H_path is not None else False device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') # ---------------------------------------- # load model # ---------------------------------------- from models.network_dpsr import MSRResNet_prior as net model = net(in_nc=n_channels + 1, out_nc=n_channels, nc=nc, nb=nb, upscale=sf, act_mode='R', upsample_mode='pixelshuffle') model.load_state_dict(torch.load(model_path), strict=False) model.eval() for k, v in model.named_parameters(): v.requires_grad = False model = model.to(device) logger.info('Model path: {:s}'.format(model_path)) number_parameters = sum(map(lambda x: x.numel(), model.parameters())) logger.info('Params number: {}'.format(number_parameters)) test_results = OrderedDict() test_results['psnr'] = [] test_results['ssim'] = [] test_results['psnr_y'] = [] test_results['ssim_y'] = [] logger.info('model_name:{}, model sigma:{}, image sigma:{}'.format( model_name, noise_level_img, noise_level_model)) logger.info(L_path) L_paths = util.get_image_paths(L_path) H_paths = util.get_image_paths(H_path) if need_H else None for idx, img in enumerate(L_paths): # ------------------------------------ # (1) img_L # ------------------------------------ img_name, ext = os.path.splitext(os.path.basename(img)) # logger.info('{:->4d}--> {:>10s}'.format(idx+1, img_name+ext)) img_L = util.imread_uint(img, n_channels=n_channels) img_L = util.uint2single(img_L) # degradation process, bicubic downsampling + Gaussian noise if need_degradation: img_L = util.modcrop(img_L, sf) img_L = util.imresize_np(img_L, 1 / sf) np.random.seed(seed=0) # for reproducibility img_L += np.random.normal(0, noise_level_img / 255., img_L.shape) util.imshow(util.single2uint(img_L), title='LR image with noise level {}'.format( noise_level_img)) if show_img else None img_L = util.single2tensor4(img_L) noise_level_map = torch.full((1, 1, img_L.size(2), img_L.size(3)), noise_level_model / 255.).type_as(img_L) img_L = torch.cat((img_L, noise_level_map), dim=1) img_L = img_L.to(device) # ------------------------------------ # (2) img_E # ------------------------------------ if not x8: img_E = model(img_L) else: img_E = utils_model.test_mode(model, img_L, mode=3, sf=sf) img_E = util.tensor2uint(img_E) if need_H: # -------------------------------- # (3) img_H # -------------------------------- img_H = util.imread_uint(H_paths[idx], n_channels=n_channels) img_H = img_H.squeeze() img_H = util.modcrop(img_H, sf) # -------------------------------- # PSNR and SSIM # -------------------------------- psnr = util.calculate_psnr(img_E, img_H, border=border) ssim = util.calculate_ssim(img_E, img_H, border=border) test_results['psnr'].append(psnr) test_results['ssim'].append(ssim) logger.info('{:s} - PSNR: {:.2f} dB; SSIM: {:.4f}.'.format( img_name + ext, psnr, ssim)) util.imshow(np.concatenate([img_E, img_H], axis=1), title='Recovered / Ground-truth') if show_img else None if np.ndim(img_H) == 3: # RGB image img_E_y = util.rgb2ycbcr(img_E, only_y=True) img_H_y = util.rgb2ycbcr(img_H, only_y=True) psnr_y = util.calculate_psnr(img_E_y, img_H_y, border=border) ssim_y = util.calculate_ssim(img_E_y, img_H_y, border=border) test_results['psnr_y'].append(psnr_y) test_results['ssim_y'].append(ssim_y) # ------------------------------------ # save results # ------------------------------------ util.imsave(img_E, os.path.join(E_path, img_name + '.png')) if need_H: ave_psnr = sum(test_results['psnr']) / len(test_results['psnr']) ave_ssim = sum(test_results['ssim']) / len(test_results['ssim']) logger.info( 'Average PSNR/SSIM(RGB) - {} - x{} --PSNR: {:.2f} dB; SSIM: {:.4f}' .format(result_name, sf, ave_psnr, ave_ssim)) if np.ndim(img_H) == 3: ave_psnr_y = sum(test_results['psnr_y']) / len( test_results['psnr_y']) ave_ssim_y = sum(test_results['ssim_y']) / len( test_results['ssim_y']) logger.info( 'Average PSNR/SSIM( Y ) - {} - x{} - PSNR: {:.2f} dB; SSIM: {:.4f}' .format(result_name, sf, ave_psnr_y, ave_ssim_y))
def main(): # ---------------------------------------- # load kernels # ---------------------------------------- PSF_grid = np.load('./data/AC254-075-A-ML-Zemax(ZMX).npz')['PSF'] PSF_grid = PSF_grid.astype(np.float32) gx, gy = PSF_grid.shape[:2] for xx in range(gx): for yy in range(gy): PSF_grid[xx, yy] = PSF_grid[xx, yy] / np.sum(PSF_grid[xx, yy], axis=(0, 1)) # ---------------------------------------- # load model # ---------------------------------------- stage = 8 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') model = net(n_iter=stage, h_nc=64, in_nc=4, out_nc=3, nc=[64, 128, 256, 512], nb=2, act_mode="R", downsample_mode='strideconv', upsample_mode="convtranspose") model_code = 'iter17000' loaded_state = torch.load( '/home/xiu/databag/deblur/models/ZEMAX/uabcnet_{}.pth'.format( model_code)) model.load_state_dict(loaded_state, strict=True) model.eval() for _, v in model.named_parameters(): v.requires_grad = False model = model.to(device) img_names = glob.glob( '/home/xiu/databag/deblur/ICCV2021/suo_image/*/AC254-075-A-ML-Zemax(ZMX).bmp' ) img_names.sort() for img_id, img_name in enumerate(img_names): img_L = cv2.imread(img_name) img_L = img_L.astype(np.float32) W, H = img_L.shape[:2] num_patch = [6, 8] #positional alpha-beta parameters for HQS ab_numpy = np.loadtxt( '/home/xiu/databag/deblur/models/ZEMAX/ab_{}.txt'.format( model_code)).astype(np.float32).reshape(gx, gy, stage * 2, 3) ab = torch.tensor(ab_numpy, device=device, requires_grad=False) #save img_L t0 = time.time() px_start = 0 py_start = 0 PSF_patch = PSF_grid[px_start:px_start + num_patch[0], py_start:py_start + num_patch[1]] #block_expand = 1 patch_L = img_L[px_start * W // gx:(px_start + num_patch[0]) * W // gx, py_start * H // gy:(py_start + num_patch[1]) * H // gy, :] p_W, p_H = patch_L.shape[:2] expand = max(PSF_grid.shape[2] // 2, p_W // 16) block_expand = expand patch_L_wrap = util_deblur.wrap_boundary_liu( patch_L, (p_W + block_expand * 2, p_H + block_expand * 2)) #centralize patch_L_wrap = np.hstack((patch_L_wrap[:, -block_expand:, :], patch_L_wrap[:, :p_H + block_expand, :])) patch_L_wrap = np.vstack((patch_L_wrap[-block_expand:, :, :], patch_L_wrap[:p_W + block_expand, :, :])) x = util.uint2single(patch_L_wrap) x = util.single2tensor4(x) k_all = [] for h_ in range(num_patch[1]): for w_ in range(num_patch[0]): k_all.append(util.single2tensor4(PSF_patch[w_, h_])) k = torch.cat(k_all, dim=0) [x, k] = [el.to(device) for el in [x, k]] ab_patch = F.softplus(ab[px_start:px_start + num_patch[0], py_start:py_start + num_patch[1]]) cd = [] for h_ in range(num_patch[1]): for w_ in range(num_patch[0]): cd.append(ab_patch[w_:w_ + 1, h_]) cd = torch.cat(cd, dim=0) x_E = model.forward_patchwise(x, k, cd, num_patch, [W // gx, H // gy]) x_E = x_E[..., block_expand:block_expand + p_W, block_expand:block_expand + p_H] patch_L = patch_L_wrap.astype(np.uint8) patch_E = util.tensor2uint(x_E) t1 = time.time() print('[{}/{}]: {} s per frame'.format(img_id, len(img_names), t1 - t0)) xk = patch_E xk = xk.astype(np.uint8) cv2.imshow('res', xk) cv2.imshow('input', patch_L.astype(np.uint8)) key = cv2.waitKey(-1) if key == ord('q'): break