def __call__(self, img, scale, flip=False, keep_ratio=True):
"""Call function of SegMapTransform."""
if keep_ratio:
img = mmcv.imrescale(img, scale, interpolation='nearest')
else:
img = mmcv.imresize(img, scale, interpolation='nearest')
if flip:
img = mmcv.imflip(img)
if self.size_divisor is not None:
img = mmcv.impad_to_multiple(img, self.size_divisor)
return img
def _pad_img(self, results):
for key in results.get('img_fields', ['img']):
if self.size is not None:
padded_img = mmcv.impad(results[key], self.size, self.pad_val)
elif self.size_divisor is not None:
padded_img = mmcv.impad_to_multiple(
results[key], self.size_divisor, pad_val=self.pad_val)
results[key] = padded_img
results['pad_shape'] = padded_img.shape
results['pad_fixed_size'] = self.size
results['pad_size_divisor'] = self.size_divisor
def _pad_img(self, results):
if self.size is not None:
padded_img = mmcv.impad(results['img'], self.size)
elif self.size_divisor is not None:
padded_img = mmcv.impad_to_multiple(results['img'],
self.size_divisor,
pad_val=self.pad_val)
results['img'] = padded_img
results['pad_shape'] = padded_img.shape
results['pad_fixed_size'] = self.size
results['pad_size_divisor'] = self.size_divisor
def _pad_img(self, results):
"""Pad images according to ``self.size``."""
if self.size is not None:
padded_img = mmcv.impad(
results['img'], shape=self.size, pad_val=self.pad_val)
elif self.size_divisor is not None:
padded_img = mmcv.impad_to_multiple(
results['img'], self.size_divisor, pad_val=self.pad_val)
results['img'] = padded_img
results['pad_shape'] = padded_img.shape
results['pad_fixed_size'] = self.size
results['pad_size_divisor'] = self.size_divisor
def __call__(self, img, scale, flip=False, v_flip=False, keep_ratio=True):
if keep_ratio:
img = mmcv.imrescale(img, scale, interpolation='nearest')
else:
img = mmcv.imresize(img, scale, interpolation='nearest')
if flip:
img = mmcv.imflip(img)
if v_flip:
img = mmcv.imflip(img, direction='vertical')
if self.size_divisor is not None:
img = mmcv.impad_to_multiple(img, self.size_divisor)
return img
def __call__(self, img, scale, flip=False):
img, scale_factor = mmcv.imrescale(img, scale, return_scale=True)
img_shape = img.shape
img = mmcv.imnormalize(img, self.mean, self.std, self.to_rgb)
if flip:
img = mmcv.imflip(img)
if self.size_divisor is not None:
img = mmcv.impad_to_multiple(img, self.size_divisor)
pad_shape = img.shape
else:
pad_shape = img_shape
img = img.transpose(2, 0, 1)
return img, img_shape, pad_shape, scale_factor
def _pad_img(self, results):
"""Pad images according to ``self.size``."""
img = results["img"]
if self.size is not None:
padded_img = mmcv.impad(img, shape=self.size, pad_val=self.pad_val)
elif self.size_divisor is not None:
padded_img = mmcv.impad_to_multiple(
img, self.size_divisor, pad_val=self.pad_val
)
results["img"] = padded_img
results["pad_shape"] = padded_img.shape
results["pad_fixed_size"] = self.size
results["pad_size_divisor"] = self.size_divisor
def _pad_img(self, results):
els = ['ref_img', 'img'] if 'ref_img' in results else ['img']
for el in els:
if self.size is not None:
padded_img = mmcv.impad(results['img'], self.size)
elif self.size_divisor is not None:
padded_img = mmcv.impad_to_multiple(results[el],
self.size_divisor,
pad_val=self.pad_val)
results[el] = padded_img
results['pad_shape'] = padded_img.shape
results['pad_fixed_size'] = self.size
results['pad_size_divisor'] = self.size_divisor
def _pad_sequence(self, results):
"""Pad images according to ``self.size``."""
if self.size is not None:
padded_sequence_imgs = [
mmcv.impad(img, shape=self.size, pad_val=self.pad_val)
for img in results['sequence_imgs']
]
elif self.size_divisor is not None:
padded_sequence_imgs = [
mmcv.impad_to_multiple(img,
self.size_divisor,
pad_val=self.pad_val)
for img in results['sequence_imgs']
]
results['sequence_imgs'] = padded_sequence_imgs
def __call__(self, img, scale, flip=False, keep_ratio=True, hsv_h=0, hsv_s=0, hsv_v=0, noisy_mode=None, blur_mode=None):
# Augment colorspace
if hsv_h+hsv_s+hsv_v > 5:
# SV augmentation by 50%
img_hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV) # hue, sat, val
H = img_hsv[:, :, 0].astype(np.float32) # hue
S = img_hsv[:, :, 1].astype(np.float32) # saturation
V = img_hsv[:, :, 2].astype(np.float32) # value
a = random.uniform(-1, 1) * hsv_h + 1
b = random.uniform(-1, 1) * hsv_s + 1
c = random.uniform(-1, 1) * hsv_v + 1
H *= a
S *= b
V *= c
img_hsv[:, :, 0] = H if a < 1 else H.clip(None, 255)
img_hsv[:, :, 1] = S if b < 1 else S.clip(None, 255)
img_hsv[:, :, 2] = V if c < 1 else V.clip(None, 255)
cv2.cvtColor(img_hsv, cv2.COLOR_HSV2BGR, dst=img)
# Add noise
if noisy_mode is not None:
img = self.add_noise(img, noisy_mode)
# Blur
if blur_mode is not None:
img = self.opencv_blur(img, blur_mode)
if keep_ratio:
img, scale_factor = mmcv.imrescale(img, scale, return_scale=True)
else:
img, w_scale, h_scale = mmcv.imresize(
img, scale, return_scale=True)
scale_factor = np.array([w_scale, h_scale, w_scale, h_scale],
dtype=np.float32)
img_shape = img.shape
img = mmcv.imnormalize(img, self.mean, self.std, self.to_rgb)
if flip:
img = mmcv.imflip(img)
if self.size_divisor is not None:
img = mmcv.impad_to_multiple(img, self.size_divisor)
pad_shape = img.shape
else:
pad_shape = img_shape
img = img.transpose(2, 0, 1)
return img, img_shape, pad_shape, scale_factor
def evaluate_single(net: STMask, im_path=None, save_path=None, idx=None):
im = mmcv.imread(im_path)
ori_shape = im.shape
im, w_scale, h_scale = mmcv.imresize(im, (640, 360), return_scale=True)
img_shape = im.shape
if cfg.backbone.transform.normalize:
im = (im - MEANS) / STD
elif cfg.backbone.transform.subtract_means:
im = (im - MEANS)
elif cfg.backbone.transform.to_float:
im = im / 255.
im = mmcv.impad_to_multiple(im, 32)
pad_shape = im.shape
im = torch.tensor(im).permute(2, 0,
1).contiguous().unsqueeze(0).float().cuda()
pad_h, pad_w = im.size()[2:4]
img_meta = {
'ori_shape': ori_shape,
'img_shape': img_shape,
'pad_shape': pad_shape
}
if idx is not None:
img_meta['frame_id'] = idx
if idx is None or idx == 0:
img_meta['is_first'] = True
else:
img_meta['is_first'] = False
preds = net(im, img_meta=[img_meta])
preds[0]['detection']['box_ids'] = torch.arange(
preds[0]['detection']['box'].size(0))
cfg.preserve_aspect_ratio = True
img_numpy = prep_display(preds[0],
im[0],
pad_h,
pad_w,
img_meta=img_meta,
img_ids=(0, idx))
if save_path is None:
plt.imshow(img_numpy)
plt.axis('off')
plt.show()
else:
cv2.imwrite(save_path, img_numpy)
def __call__(self, img, scale, flip=False, keep_ratio=True):
if keep_ratio:
img, scale_factor = mmcv.imrescale(img, scale, return_scale=True, interpolation='nearest')
else:
img, w_scale, h_scale = mmcv.imresize(
img, scale, return_scale=True)
scale_factor = np.array([w_scale, h_scale, w_scale, h_scale],
dtype=np.float32)
img_shape = img.shape
if flip:
img = mmcv.imflip(img)
if self.size_divisor is not None:
img = mmcv.impad_to_multiple(img, self.size_divisor)
pad_shape = img.shape
else:
pad_shape = img_shape
img = img[np.newaxis,:,:]
return img
def __call__(self,
img,
scale,
flip=False,
crop_info=None,
keep_ratio=True):
# image jittering
try:
img = Image.fromarray(img)
except:
print(img)
if hasattr(self, 'random_color'):
img = self.random_color(img)
if hasattr(self, 'random_contrast'):
img = self.random_contrast(img)
if hasattr(self, 'random_sharpness'):
img = self.random_sharpness(img)
img = np.array(img)
if keep_ratio:
img, scale_factor = mmcv.imrescale(img, scale, return_scale=True)
else:
img, w_scale, h_scale = mmcv.imresize(img,
scale,
return_scale=True)
scale_factor = np.array([w_scale, h_scale, w_scale, h_scale],
dtype=np.float32)
img_shape = img.shape
img = mmcv.imnormalize(img, self.mean, self.std, self.to_rgb)
if flip:
img = mmcv.imflip(img)
if crop_info is not None:
# if crop, no need to pad
cx1, cy1, cx2, cy2 = crop_info
img = img[cy1:cy2, cx1:cx2]
pad_shape = img.shape
# pad and set pad_shape
if crop_info is None and self.size_divisor is not None:
img = mmcv.impad_to_multiple(img, self.size_divisor)
pad_shape = img.shape
else:
pad_shape = img_shape
img = img.transpose(2, 0, 1)
return img, img_shape, pad_shape, scale_factor
def __call__(self, img, scale, flip=False, keep_ratio=True):
# 1. rescale/resize the image to expected size
if keep_ratio:
# Resize image while keeping the aspect ratio.
# The image will be rescaled as large as possible within the scale.
img, scale_factor = mmcv.imrescale(
img=img,
scale=scale,
return_scale=True,
interpolation='bilinear',
)
else:
# Resize image to a given size ignoring the aspect ratio.
img, w_scale, h_scale = mmcv.imresize(
img=img,
size=scale,
return_scale=True,
interpolation='bilinear',
)
scale_factor = np.array(
[w_scale, h_scale, w_scale, h_scale], dtype=np.float32
)
# 2. normalize the image
img_shape = img.shape
img = mmcv.imnormalize(img, self.mean, self.std, self.to_rgb)
# 3. flip the image (if needed)
if flip:
img = mmcv.imflip(img)
# 4. pad the image if size_divisor is not None.
# size_divisor=32 means sizes are multiplier of 32.
if self.size_divisor is not None:
img = mmcv.impad_to_multiple(img, self.size_divisor)
pad_shape = img.shape
else:
pad_shape = img_shape
# 5. transpose to (c, h, w)
img = img.transpose(2, 0, 1)
return img, img_shape, pad_shape, scale_factor
def __call__(self, img, scale, flip=False, keep_ratio=True):
if keep_ratio:
img, scale_factor = mmcv.imrescale(img, scale, return_scale=True)# scale 包括一个长边一个短边，顺序无所谓
else:# 做识别 不需要保持ratio不变，resize到 224
img, w_scale, h_scale = mmcv.imresize(
img, scale, return_scale=True)
scale_factor = np.array([w_scale, h_scale, w_scale, h_scale],
dtype=np.float32)
img_shape = img.shape
img = mmcv.imnormalize(img, self.mean, self.std, self.to_rgb)
if flip:
img = mmcv.imflip(img)
if self.size_divisor is not None:
img = mmcv.impad_to_multiple(img, self.size_divisor)# pad 边，为了每条边都能被divisor这个数字整除，比如 32
pad_shape = img.shape # pad后的图片形状
else:
pad_shape = img_shape
img = img.transpose(2, 0, 1)
return img, img_shape, pad_shape, scale_factor
def __call__(self, img, scale, flip=False, keep_ratio=True):
if keep_ratio:
img, scale_factor = mmcv.imrescale(img, scale, return_scale=True)
else:
img, w_scale, h_scale = mmcv.imresize(
img, scale, return_scale=True)
scale_factor = np.array(
[w_scale, h_scale, w_scale, h_scale], dtype=np.float32)
img_shape = img.shape
img = mmcv.imnormalize(img, self.mean, self.std, self.to_rgb) # (x - mean) / std , brg2rgb
if flip:
img = mmcv.imflip(img) # flip images
if self.size_divisor is not None:
img = mmcv.impad_to_multiple(img, self.size_divisor) # padding image to make sure divided by divisor.
pad_shape = img.shape
else:
pad_shape = img_shape
img = img.transpose(2, 0, 1) # HWC to CHW
return img, img_shape, pad_shape, scale_factor
def __call__(self, img, scale, flip=False, keep_ratio=True):
if keep_ratio:
img, scale_factor = mmcv.imrescale(img, scale, return_scale=True)
else:
img, w_scale, h_scale = mmcv.imresize(
img, scale, return_scale=True)
scale_factor = np.array([w_scale, h_scale, w_scale, h_scale],
dtype=np.float32)
img_shape = img.shape
img = mmcv.imnormalize(img, self.mean, self.std, self.to_rgb)
if flip:
img = mmcv.imflip(img)
if self.size_divisor is not None:
img = mmcv.impad_to_multiple(img, self.size_divisor)
pad_shape = img.shape
else:
pad_shape = img_shape
img = img.transpose(2, 0, 1)
return img, img_shape, pad_shape, scale_factor
def _pad_img(self, results):
for key in results.get('img_fields', ['img']):
images, img_num = [], results[key].shape[-1]
if self.size is not None:
for index in range(img_num):
images.append(
mmcv.impad(results[key][:, :, :, index], self.size,
self.pad_val))
elif self.size_divisor is not None:
for index in range(img_num):
images.append(
mmcv.impad_to_multiple(results[key][:, :, :, index],
self.size_divisor,
pad_val=self.pad_val))
images = np.stack(images, axis=-1)
results[key] = images
results['pad_shape'] = images.shape
results['pad_fixed_size'] = self.size
results['pad_size_divisor'] = self.size_divisor
def stretch(img, img_shape, gt_bboxes, gt_bboxes_8_coo, stretch_ratio,
size_divisor):
img = img.transpose(1, 2, 0)
img = img[:img_shape[0], :img_shape[1], :] # crop from padded image
height = int(img_shape[0] * stretch_ratio[1])
width = int(img_shape[1] * stretch_ratio[0])
img = cv2.resize(img, (width, height), interpolation=cv2.INTER_CUBIC)
gt_bboxes[:, [0, 2]] = gt_bboxes[:, [0, 2]] * (width / img_shape[1])
gt_bboxes[:, [1, 3]] = gt_bboxes[:, [1, 3]] * (height / img_shape[0])
gt_bboxes_8_coo[:,
0::2] = gt_bboxes_8_coo[:, 0::2] * (width / img_shape[1])
gt_bboxes_8_coo[:,
1::2] = gt_bboxes_8_coo[:, 1::2] * (width / img_shape[1])
img_shape = (height, width, 3)
img = mmcv.impad_to_multiple(img, size_divisor)
pad_shape = img.shape
img = img.transpose(2, 0, 1)
return img, img_shape, pad_shape, gt_bboxes, gt_bboxes_8_coo
def get_img(img_meta):
img = cv2.imread(img_meta[0]['filename'])
# Resize
sf = img_meta[0]['scale_factor']
img, scale_factor = mmcv.imrescale(img, sf, True)
# Normalize
m = img_meta[0]['img_norm_cfg']['mean']
s = img_meta[0]['img_norm_cfg']['std']
t = img_meta[0]['img_norm_cfg']['to_rgb']
img = mmcv.imnormalize(img, m, s, t)
# Pad
sd = 32 # size_divisor
img = mmcv.impad_to_multiple(img, 32, 0)
# H x W x C -> C x H x W and expand an dim
img = torch.from_numpy(np.transpose(img, (2, 0, 1))).expand(1, -1, -1, -1)
return img
def crop(img, img_shape, crop_ratio, gt_bboxes, gt_labels, size_divisor):
img = copy.deepcopy(img)
img_shape = copy.deepcopy(img_shape)
gt_bboxes = copy.deepcopy(gt_bboxes)
gt_labels = copy.deepcopy(gt_labels)
img = img.transpose(1, 2, 0)
height, width, _ = img_shape
left = np.random.randint(width - int(width * crop_ratio))
top = np.random.randint(height - int(height * crop_ratio))
img = img[top:top + int(height * crop_ratio),
left:left + int(width * crop_ratio), :] # crop image
img_shape = img.shape
# adjust gt bboxes to fit croped image.
i = 0
while (i < gt_bboxes.shape[0]):
if (gt_bboxes[i, 0] > left + width) or (gt_bboxes[i, 2] < left) or (
gt_bboxes[i, 1] > top + height) or (gt_bboxes[i, 3] < top):
# gt bboxes out of range of croped image
gt_bboxes = np.delete(gt_bboxes, i, axis=0)
gt_labels = np.delete(gt_labels, i)
else:
# gt bboxes in croped image or over range of some parts
if gt_bboxes[i, 0] < left:
gt_bboxes[i, 0] = left
if gt_bboxes[i, 2] > left + width - 1:
gt_bboxes[i, 2] = left + width - 1
if gt_bboxes[i, 1] < top:
gt_bboxes[i, 1] = top
if gt_bboxes[i, 3] > top + height - 1:
gt_bboxes[i, 3] = top + height - 1
i = i + 1
gt_bboxes[:, [0, 2]] = gt_bboxes[:, [0, 2]] - left
gt_bboxes[:, [1, 3]] = gt_bboxes[:, [1, 3]] - top
img = mmcv.impad_to_multiple(img, size_divisor)
pad_shape = img.shape
img = img.transpose(2, 0, 1)
return img, img_shape, pad_shape, gt_bboxes, gt_labels
def __call__(self,
img,
scale,
flip=False,
crop_info=None,
keep_ratio=True,
pad_val=0):
if keep_ratio:
img = mmcv.imrescale(img, scale, interpolation='nearest')
else:
img = mmcv.imresize(img, scale, interpolation='nearest')
if flip:
img = mmcv.imflip(img)
if crop_info is not None:
# if crop, no need to pad
cx1, cy1, cx2, cy2 = crop_info
img = img[cy1:cy2, cx1:cx2]
elif self.size_divisor is not None:
img = mmcv.impad_to_multiple(img,
self.size_divisor,
pad_val=pad_val)
return img
def processing_one_image(file_path):
img_meta = {}
img_meta['filename'] = file_path
img_meta['ori_filename'] = file_path
img_meta['flip'] = False
# 1. Read image
file_client = mmcv.FileClient(backend='disk')
img_bytes = file_client.get(file_path)
orig_img = mmcv.imfrombytes(img_bytes, flag='color') # BGR order
img_meta['ori_shape'] = orig_img.shape
# 2. Resize
test_scale = (1333, 800)
img, scale_factor = mmcv.imrescale(orig_img, test_scale, return_scale=True)
# the w_scale and h_scale has minor difference
# a real fix should be done in the mmcv.imrescale in the future
new_h, new_w = img.shape[:2]
h, w = orig_img.shape[:2]
w_scale = new_w / w
h_scale = new_h / h
scale_factor = np.array([w_scale, h_scale, w_scale, h_scale],
dtype=np.float32)
img_meta['scale_factor'] = scale_factor
img_meta['img_shape'] = img.shape
# 3. Normalize
# mean = np.array([102.9801, 115.9465, 122.7717], dtype=np.float32)
# std = np.array([1.0, 1.0, 1.0], dtype=np.float32)
mean = np.array([103.53, 116.28, 123.675], dtype=np.float32)
std = np.array([1.0, 1.0, 1.0], dtype=np.float32)
to_rgb = False
img = mmcv.imnormalize(img, mean, std, to_rgb)
img_meta['img_norm_cfg'] = dict(mean=mean, std=std, to_rgb=to_rgb)
# 4. Pad
img = mmcv.impad_to_multiple(img, divisor=32, pad_val=0)
img_meta["pad_shape"] = img.shape
# 5. ToTensor
img = torch.from_numpy(img.transpose(2, 0, 1))
return img, img_meta
def __call__(self,
img_group,
scale,
crop_history=None,
flip=False,
rotate=None,
keep_ratio=True,
dropout_prob=None,
div_255=False,
transpose=True,
stack=True):
# 1. rescale
if keep_ratio:
tuple_list = [
mmcv.imrescale(img, scale, return_scale=True)
for img in img_group
]
img_group, scale_factors = list(zip(*tuple_list))
scale_factor = scale_factors[0]
else:
tuple_list = [
mmcv.imresize(img, scale, return_scale=True)
for img in img_group
]
img_group, w_scales, h_scales = list(zip(*tuple_list))
scale_factor = np.array(
[w_scales[0], h_scales[0], w_scales[0], h_scales[0]],
dtype=np.float32)
# 2. rotate
if rotate is not None:
img_group = [mmcv.imrotate(img, rotate) for img in img_group]
# 3. crop (if necessary)
if crop_history is not None:
self.op_crop = GroupCrop(crop_history)
if self.op_crop is not None:
img_group, crop_quadruple = self.op_crop(img_group)
else:
crop_quadruple = None
img_shape = img_group[0].shape
# 4. flip
if flip:
img_group = [mmcv.imflip(img) for img in img_group]
# 5a. extra augmentation
if self.extra_augm is not None:
img_group = self.extra_augm(img_group)
# 5b. coarse dropout
if self.dropout_scale is not None and dropout_prob is not None and dropout_prob > 0.0:
dropout_mask = self._coarse_dropout_mask(img_group[0].shape,
dropout_prob,
self.dropout_scale)
img_group = [img * dropout_mask for img in img_group]
# 6a. div_255
if div_255:
img_group = [
mmcv.imnormalize(img, 0, 255, False) for img in img_group
]
# 6b. normalize
if self.mean is not None and self.std is not None:
img_group = [
mmcv.imnormalize(img, self.mean, self.std, self.to_rgb)
for img in img_group
]
elif self.to_rgb:
img_group = [mmcv.bgr2rgb(img) for img in img_group]
# 7. pad
if self.size_divisor is not None:
img_group = [
mmcv.impad_to_multiple(img, self.size_divisor)
for img in img_group
]
pad_shape = img_group[0].shape
else:
pad_shape = img_shape
# 8. transpose
if transpose:
img_group = [img.transpose((2, 0, 1)) for img in img_group]
# 9. stack into numpy.array
if stack:
img_group = np.stack(img_group, axis=0)
return img_group, img_shape, pad_shape, scale_factor, crop_quadruple
def __call__(self, results):
# if self.resize_crop or self.rescale_crop:
# img_group, crop_quadruple = self.op_crop(img_group)
# img_shape = img_group[0].shape
# scale_factor = None
# else:
# 1. rescale
# if keep_ratio:
img_group = results['img_group']
tuple_list = [
mmcv.imrescale(img, self.scale, return_scale=True)
for img in img_group
]
img_group, scale_factors = list(zip(*tuple_list))
scale_factor = scale_factors[0]
# else:
# tuple_list = [mmcv.imresize(
# img, scale, return_scale=True) for img in img_group]
# img_group, w_scales, h_scales = list(zip(*tuple_list))
# scale_factor = np.array([w_scales[0], h_scales[0],
# w_scales[0], h_scales[0]],
# dtype=np.float32)
# 2. crop (if necessary)
# if crop_history is not None:
# self.op_crop = GroupCrop(crop_history)
# if self.op_crop is not None:
# img_group, crop_quadruple = self.op_crop(
# img_group, is_flow=is_flow)
# else:
crop_quadruple = None
img_shape = img_group[0].shape
# # 3. flip
# if flip:
# img_group = [mmcv.imflip(img) for img in img_group]
# if is_flow:
# for i in range(0, len(img_group), 2):
# img_group[i] = mmcv.iminvert(img_group[i])
# # 4a. div_255
# if div_255:
# img_group = [mmcv.imnormalize(img, 0, 255, False)
# for img in img_group]
# 4. normalize
img_group = [
mmcv.imnormalize(img, self.mean, self.std, self.to_rgb)
for img in img_group
]
# 5. pad
if self.size_divisor is not None:
img_group = [
mmcv.impad_to_multiple(img, self.size_divisor)
for img in img_group
]
pad_shape = img_group[0].shape
else:
pad_shape = img_shape
# if is_flow:
# assert len(img_group[0].shape) == 2
# img_group = [np.stack((flow_x, flow_y), axis=2)
# for flow_x, flow_y in zip(
# img_group[0::2], img_group[1::2])]
# 6. transpose
img_group = [img.transpose(2, 0, 1) for img in img_group]
# Stack into numpy.array
img_group = np.stack(img_group, axis=0)
img_group = to_tensor(img_group)
img_group = np.transpose(img_group, (1, 0, 2, 3))
results['img_group_0'] = [img_group] #img_group
results['img_shape'] = img_shape
results['pad_shape'] = pad_shape
results['scale_factor'] = scale_factor
results['crop_quadruple'] = crop_quadruple
results['flip'] = False
return results
def __call__(self,
img_group,
scale,
crop_history=None,
flip=False,
keep_ratio=True,
div_255=False,
is_flow=False,
interpolation='bilinear',
normalize=True,
more_aug=False):
# 1. rescale
if keep_ratio:
tuple_list = [
mmcv.imrescale(img,
scale,
return_scale=True,
interpolation=interpolation)
for img in img_group
]
img_group, scale_factors = list(zip(*tuple_list))
scale_factor = scale_factors[0]
else:
tuple_list = [
mmcv.imresize(img,
scale,
return_scale=True,
interpolation=interpolation) for img in img_group
]
img_group, w_scales, h_scales = list(zip(*tuple_list))
scale_factor = np.array(
[w_scales[0], h_scales[0], w_scales[0], h_scales[0]],
dtype=np.float32)
# 2. crop (if necessary)
if crop_history is not None:
self.op_crop = GroupCrop(crop_history,
input_size=self.crop_size,
resize=True)
if self.op_crop is not None and isinstance(
self.op_crop, (GroupCrop, GroupMultiScaleCrop)):
img_group, crop_quadruple = self.op_crop(
img_group, is_flow=is_flow, interpolation=interpolation)
elif self.op_crop is not None:
img_group, crop_quadruple = self.op_crop(img_group,
is_flow=is_flow)
else:
crop_quadruple = None
img_shape = img_group[0].shape
if more_aug:
seq = iaa.Sequential([
iaa.GaussianBlur(sigma=np.random.uniform(0, 5)),
iaa.AdditiveGaussianNoise(loc=0,
scale=(0.0, 0.05 * 255),
per_channel=0.5),
])
img_group = seq(images=np.array(img_group))
# 3. flip
if flip:
img_group = [mmcv.imflip(img) for img in img_group]
if is_flow:
for i in range(0, len(img_group), 2):
img_group[i] = mmcv.iminvert(img_group[i])
# 4a. div_255
if div_255:
img_group = [
mmcv.imnormalize(img, 0, 255, False) for img in img_group
]
# 4. normalize
if normalize:
img_group = [
mmcv.imnormalize(img, self.mean, self.std, self.to_rgb)
for img in img_group
]
# 5. pad
if self.size_divisor is not None:
img_group = [
mmcv.impad_to_multiple(img, self.size_divisor)
for img in img_group
]
pad_shape = img_group[0].shape
else:
pad_shape = img_shape
if is_flow:
assert len(img_group[0].shape) == 2
img_group = [
np.stack((flow_x, flow_y), axis=2)
for flow_x, flow_y in zip(img_group[0::2], img_group[1::2])
]
# 6. transpose
if len(img_shape) == 2:
img_group = [img[:, :, np.newaxis] for img in img_group]
img_group = [img.transpose(2, 0, 1) for img in img_group]
# Stack into numpy.array
img_group = np.stack(img_group, axis=0)
return img_group, img_shape, pad_shape, scale_factor, crop_quadruple
def __call__(self,
img_group,
scale,
crop_history=None,
flip=False,
keep_ratio=True,
div_255=False,
is_flow=False):
if self.resize_crop or self.rescale_crop:
img_group, crop_quadruple = self.op_crop(img_group)
img_shape = img_group[0].shape
scale_factor = None
else:
# 1. rescale
if keep_ratio:
tuple_list = [
mmcv.imrescale(img, scale, return_scale=True)
for img in img_group
]
img_group, scale_factors = list(zip(*tuple_list))
scale_factor = scale_factors[0]
else:
tuple_list = [
mmcv.imresize(img, scale, return_scale=True)
for img in img_group
]
img_group, w_scales, h_scales = list(zip(*tuple_list))
scale_factor = np.array(
[w_scales[0], h_scales[0], w_scales[0], h_scales[0]],
dtype=np.float32)
# 2. crop (if necessary)
if crop_history is not None:
self.op_crop = GroupCrop(crop_history)
if self.op_crop is not None:
img_group, crop_quadruple = self.op_crop(img_group,
is_flow=is_flow)
else:
crop_quadruple = None
img_shape = img_group[0].shape
# 3. flip
if flip:
img_group = [mmcv.imflip(img) for img in img_group]
if is_flow:
for i in range(0, len(img_group), 2):
img_group[i] = mmcv.iminvert(img_group[i])
# 4a. div_255
if div_255:
img_group = [
mmcv.imnormalize(img, 0, 255, False) for img in img_group
]
# 4. normalize
img_group = [
mmcv.imnormalize(img, self.mean, self.std, self.to_rgb)
for img in img_group
]
# 5. pad
if self.size_divisor is not None:
img_group = [
mmcv.impad_to_multiple(img, self.size_divisor)
for img in img_group
]
pad_shape = img_group[0].shape
else:
pad_shape = img_shape
if is_flow:
assert len(img_group[0].shape) == 2
img_group = [
np.stack((flow_x, flow_y), axis=2)
for flow_x, flow_y in zip(img_group[0::2], img_group[1::2])
]
# 6. transpose
img_group = [img.transpose(2, 0, 1) for img in img_group]
# Stack into numpy.array
img_group = np.stack(img_group, axis=0)
return img_group, img_shape, pad_shape, scale_factor, crop_quadruple
# flip the image vertically mmcv.imflip(img, direction='vertical') # ============Crop============= # crop the region (10, 10, 100, 120) bboxes = np.array([10, 10, 100, 120]) patch = mmcv.imcrop(img, bboxes) # crop two regions (10, 10, 100, 120) and (0, 0, 50, 50) bboxes = np.array([[10, 10, 100, 120], [0, 0, 50, 50]]) patches = mmcv.imcrop(img, bboxes) # crop two regions, and rescale the patches by 1.2x patches = mmcv.imcrop(img, bboxes, scale_ratio=1.2) # =============Padding============= # pad the image to (1000, 1200) with all zeros img_ = mmcv.impad(img, (1000, 1200), pad_val=0) # pad the image to (1000, 1200) with different values for three channels. img_ = mmcv.impad(img, (1000, 1200), pad_val=[100, 50, 200]) # pad an image so that each edge is a multiple of some value. img_ = mmcv.impad_to_multiple(img, 32)
def __call__(self,
img_group,
scale,
crop_history=None,
flip=False,
keep_ratio=True,
div_255=False,
is_flow=False,
image_name=None):
if self.resize_crop or self.rescale_crop:
if self.afterdetect_resize:
img_group = [
cv2.resize(img, self.afterdetect_resize)
for img in img_group
]
img_group, crop_quadruple = self.op_crop(img_group)
img_shape = img_group[0].shape
scale_factor = None
else:
# 1. rescale
if keep_ratio:
tuple_list = [
mmcv.imrescale(img, scale, return_scale=True)
for img in img_group
]
img_group, scale_factors = list(zip(*tuple_list))
scale_factor = scale_factors[0]
else:
tuple_list = [
mmcv.imresize(img, scale, return_scale=True)
for img in img_group
]
img_group, w_scales, h_scales = list(zip(*tuple_list))
scale_factor = np.array(
[w_scales[0], h_scales[0], w_scales[0], h_scales[0]],
dtype=np.float32)
if self.pre_mean_volume is not None:
volume_len = self.pre_mean_volume.shape[0]
img_group = [
img - self.pre_mean_volume[i % volume_len, ...]
for i, img in enumerate(img_group)
]
# 2. crop (if necessary)
if crop_history is not None:
self.op_crop = GroupCrop(crop_history)
if self.op_crop is not None:
img_group, crop_quadruple = self.op_crop(img_group,
is_flow=is_flow)
else:
crop_quadruple = None
img_shape = img_group[0].shape
# import matplotlib.pyplot as plt
# import os
# for i in range(len(img_group)):
# plt.imshow(img_group[i])
# save_dir = os.path.join("show_fig",image_name.split("/")[-1])
# if not os.path.exists(save_dir):
# os.mkdir(save_dir)
# plt.savefig(os.path.join(save_dir,"{}".format(i)))
# 3. flip
if flip:
img_group = [mmcv.imflip(img) for img in img_group]
if is_flow:
for i in range(0, len(img_group), 2):
img_group[i] = mmcv.iminvert(img_group[i])
# 4a. div_255
if div_255:
img_group = [
mmcv.imnormalize(img, 0, 255, False) for img in img_group
]
# 4. normalize
img_group = [
mmcv.imnormalize(img, self.mean, self.std, self.to_rgb)
for img in img_group
]
# 5. pad
if self.size_divisor is not None:
img_group = [
mmcv.impad_to_multiple(img, self.size_divisor)
for img in img_group
]
pad_shape = img_group[0].shape
else:
pad_shape = img_shape
if is_flow:
assert len(img_group[0].shape) == 2
img_group = [
np.stack((flow_x, flow_y), axis=2)
for flow_x, flow_y in zip(img_group[0::2], img_group[1::2])
]
# 6. transpose
img_group = [img.transpose(2, 0, 1) for img in img_group]
# Stack into numpy.array
img_group = np.stack(img_group, axis=0)
return img_group, img_shape, pad_shape, scale_factor, crop_quadruple