def test(args, test_loader, model):
"""
args:
test_loader: loaded for test dataset
model: model
return: class IoU and mean IoU
"""
# evaluation or test mode
model.eval()
total_batches = len(test_loader)
metric = SegmentationMetric(numClass=args.classes)
pbar = tqdm(iterable=enumerate(test_loader),
total=total_batches,
desc='Valing')
for i, (input, gt, size, name) in pbar:
with torch.no_grad():
input_var = Variable(input).cuda()
output = model(input_var)
torch.cuda.synchronize()
output = output.cpu().data[0].numpy()
output = output.transpose(1, 2, 0)
output = np.asarray(np.argmax(output, axis=2), dtype=np.uint8)
gt = np.asarray(gt[0], dtype=np.uint8)
# 计算miou
metric.addBatch(imgPredict=output.flatten(), imgLabel=gt.flatten())
# save the predicted image
if args.save:
save_predict(output,
gt,
name[0],
args.dataset,
args.save_seg_dir,
output_grey=False,
output_color=True,
gt_color=True)
pa = metric.pixelAccuracy()
cpa = metric.classPixelAccuracy()
mpa = metric.meanPixelAccuracy()
Miou, PerMiou_set = metric.meanIntersectionOverUnion()
FWIoU = metric.Frequency_Weighted_Intersection_over_Union()
return Miou, PerMiou_set, FWIoU, pa, mpa
def predict(args, test_loader, model):
"""
args:
test_loader: loaded for test dataset, for those that do not provide label on the test set
model: model
return: class IoU and mean IoU
"""
# evaluation or test mode
model.eval()
total_batches = len(test_loader)
metric = SegmentationMetric(numClass=args.classes)
pbar = tqdm(iterable=enumerate(test_loader), total=total_batches, desc='Valing')
for i, (input, gt, size, name) in pbar:
with torch.no_grad():
input_var = input.cuda()
output = model(input_var)
torch.cuda.synchronize()
output = output.cpu().data[0].numpy()
output = output.transpose(1, 2, 0)
output = np.asarray(np.argmax(output, axis=2), dtype=np.uint8)
gt = np.asarray(gt[0], dtype=np.uint8)
# 计算miou
metric.addBatch(imgPredict=output.flatten(), imgLabel=gt.flatten())
# save the predicted image
save_predict(output, gt, name[0], args.dataset, args.save_seg_dir,
output_grey=False, output_color=True, gt_color=True)
pa = metric.pixelAccuracy()
cpa = metric.classPixelAccuracy()
mpa = metric.meanPixelAccuracy()
Miou, PerMiou_set = metric.meanIntersectionOverUnion()
FWIoU = metric.Frequency_Weighted_Intersection_over_Union()
print('miou {}\nclass iou {}'.format(Miou, PerMiou_set))
result = args.save_seg_dir + '/results.txt'
with open(result, 'w') as f:
f.write(str(Miou))
f.write('\n{}'.format(PerMiou_set))
def val(args, val_loader, criteria, model):
"""
args:
val_loader: loaded for validation dataset
model: model
return: mean IoU and IoU class
"""
# evaluation mode
model.eval()
total_batches = len(val_loader)
val_loss = []
metric = SegmentationMetric(args.classes)
pbar = tqdm(iterable=enumerate(val_loader),
total=total_batches,
desc='Val')
for iteration, (input, label, size, name) in pbar:
with torch.no_grad():
input_var = input.cuda().float()
output = model(input_var)
if type(output) is tuple:
output = output[0]
loss = criteria(output, label.long().cuda())
val_loss.append(loss)
output = output.cpu().data[0].numpy()
gt = np.asarray(label[0].numpy(), dtype=np.uint8)
output = output.transpose(1, 2, 0)
output = np.asarray(np.argmax(output, axis=2), dtype=np.uint8)
# 计算miou
metric.addBatch(imgPredict=output.flatten(), imgLabel=gt.flatten())
val_loss = sum(val_loss) / len(val_loss)
pa = metric.pixelAccuracy()
cpa = metric.classPixelAccuracy()
mpa = metric.meanPixelAccuracy()
Miou, PerMiou_set = metric.meanIntersectionOverUnion()
FWIoU = metric.Frequency_Weighted_Intersection_over_Union()
return val_loss, FWIoU, Miou, PerMiou_set
def predict_sliding(args, net, image, tile_size, classes):
total_batches = len(image)
metric = SegmentationMetric(args.classes) # args.classes表示有args.classes个分类
pbar = tqdm(iterable=enumerate(image),
total=total_batches,
desc='Predicting')
for i, (input, gt, size, name) in pbar:
image_size = input.shape # (1,3,3328,3072)
overlap = 1 / 3 # 每次滑动的覆盖率为1/3
# print(image_size, tile_size)
stride = ceil(
tile_size[0] *
(1 - overlap)) # 滑动步长:512*(1-1/3) = 513 512*(1-1/3)= 342
tile_rows = int(ceil((image_size[2] - tile_size[0]) / stride) +
1) # 行滑动步数:(3072-512)/342+1=9
tile_cols = int(ceil((image_size[3] - tile_size[1]) / stride) +
1) # 列滑动步数:(3328-512)/342+1=10
full_probs = np.zeros((image_size[2], image_size[3],
classes)) # 初始化全概率矩阵shape(3072,3328,3)
count_predictions = np.zeros((image_size[2], image_size[3],
classes)) # 初始化计数矩阵shape(3072,3328,3)
for row in range(tile_rows): # row = 0,1 0,1,2,3,4,5,6,7,8
for col in range(
tile_cols): # col = 0,1,2,3 0,1,2,3,4,5,6,7,8,9
x1 = int(col * stride) # 起始位置x1 = 0 * 513 = 0 0*342
y1 = int(row * stride) # y1 = 0 * 513 = 0 0*342
x2 = min(x1 + tile_size[1],
image_size[3]) # 末位置x2 = min(0+512, 3328)
y2 = min(y1 + tile_size[0],
image_size[2]) # y2 = min(0+512, 3072)
x1 = max(int(x2 - tile_size[1]),
0) # 重新校准起始位置x1 = max(512-512, 0)
y1 = max(int(y2 - tile_size[0]), 0) # y1 = max(512-512, 0)
img = input[:, :, y1:y2,
x1:x2] # 滑动窗口对应的图像 imge[:, :, 0:512, 0:512]
padded_img = pad_image(img,
tile_size) # padding 确保扣下来的图像为512*512
# plt.imshow(padded_img)
# plt.show()
# 将扣下来的部分传入网络，网络输出概率图。
with torch.no_grad():
input_var = torch.from_numpy(padded_img).cuda().float()
padded_prediction = net(input_var)
if type(padded_prediction) is tuple:
padded_prediction = padded_prediction[0]
torch.cuda.synchronize()
if isinstance(padded_prediction, list):
padded_prediction = padded_prediction[
0] # shape(1,3,512,512)
padded_prediction = padded_prediction.cpu().data[0].numpy(
).transpose(1, 2, 0) # 通道位置变换(512,512,3)
prediction = padded_prediction[
0:img.shape[2],
0:img.shape[3], :] # 扣下相应面积 shape(512,512,3)
count_predictions[y1:y2, x1:x2] += 1 # 窗口区域内的计数矩阵加1
full_probs[y1:y2, x1:x2] += prediction # 窗口区域内的全概率矩阵叠加预测结果
# average the predictions in the overlapping regions
full_probs /= count_predictions # 全概率矩阵 除以 计数矩阵 即得 平均概率
# visualize normalization Weights
# plt.imshow(np.mean(count_predictions, axis=2))
# plt.show()
full_probs = np.asarray(np.argmax(full_probs, axis=2), dtype=np.uint8)
'''设置输出原图和预测图片的颜色灰度还是彩色'''
gt = gt[0].numpy()
# 计算miou
metric.addBatch(full_probs, gt)
save_predict(full_probs,
gt,
name[0],
args.dataset,
args.save_seg_dir,
output_grey=False,
output_color=True,
gt_color=True)
pa = metric.pixelAccuracy()
cpa = metric.classPixelAccuracy()
mpa = metric.meanPixelAccuracy()
Miou, PerMiou_set = metric.meanIntersectionOverUnion()
FWIoU = metric.Frequency_Weighted_Intersection_over_Union()
print('miou {}\nclass iou {}'.format(Miou, PerMiou_set))
result = args.save_seg_dir + '/results.txt'
with open(result, 'w') as f:
f.write(str(Miou))
f.write('\n{}'.format(PerMiou_set))
def test(args, test_loader, model):
"""
args:
test_loader: loaded for test dataset
model: model
return: class IoU and mean IoU
"""
# evaluation or test mode
model.eval()
total_batches = len(test_loader)
Miou_list = []
Iou_list = []
Pa_list = []
Mpa_list = []
Fmiou_list = []
pbar = tqdm(iterable=enumerate(test_loader),
total=total_batches,
desc='Valing')
for i, (input, gt, size, name) in pbar:
with torch.no_grad():
input_var = Variable(input).cuda()
start_time = time.time()
output = model(input_var)
torch.cuda.synchronize()
time_taken = time.time() - start_time
pbar.set_postfix(cost_time='%.3f' % time_taken)
output = output.cpu().data[0].numpy()
output = output.transpose(1, 2, 0)
output = np.asarray(np.argmax(output, axis=2), dtype=np.uint8)
gt = np.asarray(gt[0], dtype=np.uint8)
# 计算miou
metric = SegmentationMetric(numClass=args.classes)
metric.addBatch(imgPredict=output, imgLabel=gt)
miou, iou = metric.meanIntersectionOverUnion()
fmiou = metric.Frequency_Weighted_Intersection_over_Union()
pa = metric.pixelAccuracy()
mpa = metric.meanPixelAccuracy()
Miou_list.append(miou)
Fmiou_list.append(fmiou)
Pa_list.append(pa)
Mpa_list.append(mpa)
iou = np.array(iou)
Iou_list.append(iou)
# save the predicted image
if args.save:
save_predict(output,
gt,
name[0],
args.dataset,
args.save_seg_dir,
output_grey=False,
output_color=True,
gt_color=True)
miou = np.mean(Miou_list)
fmiou = np.mean(Fmiou_list)
pa = np.mean(Pa_list)
mpa = np.mean(Mpa_list)
Iou_list = np.asarray(Iou_list)
iou = np.mean(Iou_list, axis=0)
cls_iu = dict(zip(range(args.classes), iou))
return miou, cls_iu, fmiou, pa, mpa