def calculate_psnr(img1,
img2,
crop_border,
input_order='HWC',
test_y_channel=False):
"""Calculate PSNR (Peak Signal-to-Noise Ratio).
Ref: https://en.wikipedia.org/wiki/Peak_signal-to-noise_ratio
Args:
img1 (ndarray/tensor): Images with range [0, 255]/[0, 1].
img2 (ndarray/tensor): Images with range [0, 255]/[0, 1].
crop_border (int): Cropped pixels in each edge of an image. These
pixels are not involved in the PSNR calculation.
input_order (str): Whether the input order is 'HWC' or 'CHW'.
Default: 'HWC'.
test_y_channel (bool): Test on Y channel of YCbCr. Default: False.
Returns:
float: psnr result.
"""
assert img1.shape == img2.shape, (
f'Image shapes are differnet: {img1.shape}, {img2.shape}.')
if input_order not in ['HWC', 'CHW']:
raise ValueError(
f'Wrong input_order {input_order}. Supported input_orders are '
'"HWC" and "CHW"')
if type(img1) == torch.Tensor:
if len(img1.shape) == 4:
img1 = img1.squeeze(0)
img1 = img1.detach().cpu().numpy().transpose(1, 2, 0)
if type(img2) == torch.Tensor:
if len(img2.shape) == 4:
img2 = img2.squeeze(0)
img2 = img2.detach().cpu().numpy().transpose(1, 2, 0)
img1 = reorder_image(img1, input_order=input_order)
img2 = reorder_image(img2, input_order=input_order)
img1 = img1.astype(np.float64)
img2 = img2.astype(np.float64)
if crop_border != 0:
img1 = img1[crop_border:-crop_border, crop_border:-crop_border, ...]
img2 = img2[crop_border:-crop_border, crop_border:-crop_border, ...]
if test_y_channel:
img1 = to_y_channel(img1)
img2 = to_y_channel(img2)
mse = np.mean((img1 - img2)**2)
if mse == 0:
return float('inf')
max_value = 1. if img1.max() <= 1 else 255.
return 20. * np.log10(max_value / np.sqrt(mse))
def calculate_ssim(img,
img2,
crop_border,
input_order='HWC',
test_y_channel=False,
**kwargs):
"""Calculate SSIM (structural similarity).
Ref:
Image quality assessment: From error visibility to structural similarity
The results are the same as that of the official released MATLAB code in
https://ece.uwaterloo.ca/~z70wang/research/ssim/.
For three-channel images, SSIM is calculated for each channel and then
averaged.
Args:
img (ndarray): Images with range [0, 255].
img2 (ndarray): Images with range [0, 255].
crop_border (int): Cropped pixels in each edge of an image. These pixels are not involved in the calculation.
input_order (str): Whether the input order is 'HWC' or 'CHW'.
Default: 'HWC'.
test_y_channel (bool): Test on Y channel of YCbCr. Default: False.
Returns:
float: SSIM result.
"""
assert img.shape == img2.shape, (
f'Image shapes are different: {img.shape}, {img2.shape}.')
if input_order not in ['HWC', 'CHW']:
raise ValueError(
f'Wrong input_order {input_order}. Supported input_orders are "HWC" and "CHW"'
)
img = reorder_image(img, input_order=input_order)
img2 = reorder_image(img2, input_order=input_order)
if crop_border != 0:
img = img[crop_border:-crop_border, crop_border:-crop_border, ...]
img2 = img2[crop_border:-crop_border, crop_border:-crop_border, ...]
if test_y_channel:
img = to_y_channel(img)
img2 = to_y_channel(img2)
img = img.astype(np.float64)
img2 = img2.astype(np.float64)
ssims = []
for i in range(img.shape[2]):
ssims.append(_ssim(img[..., i], img2[..., i]))
return np.array(ssims).mean()
def calculate_lpips(img1, img2, crop_border, input_order='HWC'):
"""Calculate LPIPS metric.
We use the official params estimated from the pristine dataset.
We use the recommended block size (96, 96) without overlaps.
Args:
img (ndarray): Input image whose quality needs to be computed.
The input image must be in range [0, 255] with float/int type.
The input_order of image can be 'HW' or 'HWC' or 'CHW'. (BGR order)
If the input order is 'HWC' or 'CHW', it will be converted to gray
or Y (of YCbCr) image according to the ``convert_to`` argument.
crop_border (int): Cropped pixels in each edge of an image. These
pixels are not involved in the metric calculation.
input_order (str): Whether the input order is 'HW', 'HWC' or 'CHW'.
Default: 'HWC'.
Returns:
float: LPIPS result.
"""
assert img1.shape == img2.shape, (
f'Image shapes are differnet: {img1.shape}, {img2.shape}.')
if input_order not in ['HWC', 'CHW']:
raise ValueError(
f'Wrong input_order {input_order}. Supported input_orders are '
'"HWC" and "CHW"')
img1 = reorder_image(img1, input_order=input_order)
img2 = reorder_image(img2, input_order=input_order)
img1 = img1.astype(np.float64)
img2 = img2.astype(np.float64)
if crop_border != 0:
img1 = img1[crop_border:-crop_border, crop_border:-crop_border, ...]
img2 = img2[crop_border:-crop_border, crop_border:-crop_border, ...]
img1 = img1.astype(np.float32)
img2 = img2.astype(np.float32)
img1, img2 = totensor([img1, img2], bgr2rgb=False, float32=True)
img1 = img1.unsqueeze(0)
img2 = img2.unsqueeze(0)
# image should be RGB, IMPORTANT: normalized to [-1,1]
img1 = (img1 / 255. - 0.5) * 2
img2 = (img2 / 255. - 0.5) * 2
loss_fn_alex = lpips.LPIPS(net='alex',
verbose=False) # best forward scores
metric = loss_fn_alex(img1, img2).squeeze(0).float().detach().cpu().numpy()
return metric.mean()
def calculate_psnr(img,
img2,
crop_border,
input_order='HWC',
test_y_channel=False,
**kwargs):
"""Calculate PSNR (Peak Signal-to-Noise Ratio).
Ref: https://en.wikipedia.org/wiki/Peak_signal-to-noise_ratio
Args:
img (ndarray): Images with range [0, 255].
img2 (ndarray): Images with range [0, 255].
crop_border (int): Cropped pixels in each edge of an image. These
pixels are not involved in the PSNR calculation.
input_order (str): Whether the input order is 'HWC' or 'CHW'.
Default: 'HWC'.
test_y_channel (bool): Test on Y channel of YCbCr. Default: False.
Returns:
float: psnr result.
"""
assert img.shape == img2.shape, (
f'Image shapes are different: {img.shape}, {img2.shape}.')
if input_order not in ['HWC', 'CHW']:
raise ValueError(
f'Wrong input_order {input_order}. Supported input_orders are "HWC" and "CHW"'
)
img = reorder_image(img, input_order=input_order)
img2 = reorder_image(img2, input_order=input_order)
img = img.astype(np.float64)
img2 = img2.astype(np.float64)
if crop_border != 0:
img = img[crop_border:-crop_border, crop_border:-crop_border, ...]
img2 = img2[crop_border:-crop_border, crop_border:-crop_border, ...]
if test_y_channel:
img = to_y_channel(img)
img2 = to_y_channel(img2)
mse = np.mean((img - img2)**2)
if mse == 0:
return float('inf')
return 20. * np.log10(255. / np.sqrt(mse))
def calculate_niqe(img,
crop_border,
input_order='HWC',
convert_to='y',
**kwargs):
"""Calculate NIQE (Natural Image Quality Evaluator) metric.
Ref: Making a "Completely Blind" Image Quality Analyzer.
This implementation could produce almost the same results as the official
MATLAB codes: http://live.ece.utexas.edu/research/quality/niqe_release.zip
> MATLAB R2021a result for tests/data/baboon.png: 5.72957338 (5.7296)
> Our re-implementation result for tests/data/baboon.png: 5.7295763 (5.7296)
We use the official params estimated from the pristine dataset.
We use the recommended block size (96, 96) without overlaps.
Args:
img (ndarray): Input image whose quality needs to be computed.
The input image must be in range [0, 255] with float/int type.
The input_order of image can be 'HW' or 'HWC' or 'CHW'. (BGR order)
If the input order is 'HWC' or 'CHW', it will be converted to gray
or Y (of YCbCr) image according to the ``convert_to`` argument.
crop_border (int): Cropped pixels in each edge of an image. These
pixels are not involved in the metric calculation.
input_order (str): Whether the input order is 'HW', 'HWC' or 'CHW'.
Default: 'HWC'.
convert_to (str): Whether converted to 'y' (of MATLAB YCbCr) or 'gray'.
Default: 'y'.
Returns:
float: NIQE result.
"""
ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
# we use the official params estimated from the pristine dataset.
niqe_pris_params = np.load(os.path.join(ROOT_DIR, 'niqe_pris_params.npz'))
mu_pris_param = niqe_pris_params['mu_pris_param']
cov_pris_param = niqe_pris_params['cov_pris_param']
gaussian_window = niqe_pris_params['gaussian_window']
img = img.astype(np.float32)
if input_order != 'HW':
img = reorder_image(img, input_order=input_order)
if convert_to == 'y':
img = to_y_channel(img)
elif convert_to == 'gray':
img = cv2.cvtColor(img / 255., cv2.COLOR_BGR2GRAY) * 255.
img = np.squeeze(img)
if crop_border != 0:
img = img[crop_border:-crop_border, crop_border:-crop_border]
# round is necessary for being consistent with MATLAB's result
img = img.round()
niqe_result = niqe(img, mu_pris_param, cov_pris_param, gaussian_window)
return niqe_result
def calculate_ssim(img1,
img2,
crop_border,
input_order='HWC',
test_y_channel=False):
"""Calculate SSIM (structural similarity).
Ref:
Image quality assessment: From error visibility to structural similarity
The results are the same as that of the official released MATLAB code in
https://ece.uwaterloo.ca/~z70wang/research/ssim/.
For three-channel images, SSIM is calculated for each channel and then
averaged.
Args:
img1 (ndarray): Images with range [0, 255].
img2 (ndarray): Images with range [0, 255].
crop_border (int): Cropped pixels in each edge of an image. These
pixels are not involved in the SSIM calculation.
input_order (str): Whether the input order is 'HWC' or 'CHW'.
Default: 'HWC'.
test_y_channel (bool): Test on Y channel of YCbCr. Default: False.
Returns:
float: ssim result.
"""
assert img1.shape == img2.shape, (
f'Image shapes are differnet: {img1.shape}, {img2.shape}.')
if input_order not in ['HWC', 'CHW']:
raise ValueError(
f'Wrong input_order {input_order}. Supported input_orders are '
'"HWC" and "CHW"')
if type(img1) == torch.Tensor:
if len(img1.shape) == 4:
img1 = img1.squeeze(0)
img1 = img1.detach().cpu().numpy().transpose(1, 2, 0)
if type(img2) == torch.Tensor:
if len(img2.shape) == 4:
img2 = img2.squeeze(0)
img2 = img2.detach().cpu().numpy().transpose(1, 2, 0)
img1 = reorder_image(img1, input_order=input_order)
img2 = reorder_image(img2, input_order=input_order)
img1 = img1.astype(np.float64)
img2 = img2.astype(np.float64)
if crop_border != 0:
img1 = img1[crop_border:-crop_border, crop_border:-crop_border, ...]
img2 = img2[crop_border:-crop_border, crop_border:-crop_border, ...]
if test_y_channel:
img1 = to_y_channel(img1)
img2 = to_y_channel(img2)
return _ssim_cly(img1[..., 0], img2[..., 0])
ssims = []
# ssims_before = []
# skimage_before = skimage.metrics.structural_similarity(img1, img2, data_range=255., multichannel=True)
# print('.._skimage',
# skimage.metrics.structural_similarity(img1, img2, data_range=255., multichannel=True))
max_value = 1 if img1.max() <= 1 else 255
with torch.no_grad():
final_ssim = _ssim_3d(img1, img2, max_value)
ssims.append(final_ssim)
# for i in range(img1.shape[2]):
# ssims_before.append(_ssim(img1, img2))
# print('..ssim mean , new {:.4f} and before {:.4f} .... skimage before {:.4f}'.format(np.array(ssims).mean(), np.array(ssims_before).mean(), skimage_before))
# ssims.append(skimage.metrics.structural_similarity(img1[..., i], img2[..., i], multichannel=False))
return np.array(ssims).mean()
