o .wi U@sddlmZddlmZmZmZddlZddlmZddlm Z ddl m Z ddl mZmZmZddlmZdd lmZd ed ed eeeffd dZ        d6d ed ededeeeefdeeeefdeeeeeeffdedededed eeeeefffddZ d7d ed!e d"d efd#d$Z         d8d ed ededeeeefdeeeefd!e d"deeeeeeffdedededed eeeeefffd%d&Z       d9d ed ededeeeefdeeeefdeeeeeeffdeded'ee d(d eeeffd)d*Z       + d:d ed ededeeeefdeeeefdeeeeeeffdeded,eeeeeeefeed-ffd'ee d(d efd.d/Z d7d0ed!e d"d efd1d2Z        + 3d;d ed ededeeeefdeeeefd!e d"deeeeeeffdeded,eed-fd'ee d(d efd4d5Z dS)<)Sequence)ListOptionalUnionN)Tensor) functional)Literal)_gaussian_kernel_2d_gaussian_kernel_3d_reflection_pad_3d)_check_same_shapereducepredstargetreturncCsR|j|jkr ||j}t||t|jdvr%td|jd|jd||fS)zUpdate and returns variables required to compute Structural Similarity Index Measure. Args: preds: Predicted tensor target: Ground truth tensor )zMExpected `preds` and `target` to have BxCxHxW or BxCxDxHxW shape. Got preds: z and target: .)dtypetor lenshape ValueError)rrr_/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/torchmetrics/functional/image/ssim.py_ssim_check_inputss   rT? {Gz?Q?Fgaussian_kernelsigma kernel_size data_rangek1k2return_full_imagereturn_contrast_sensitivityc "Cs|jdk} t|ts| rd|gnd|g}t|ts'| r"d|gnd|g}t|t|jdkrAtdt|dt|jt|dvrPtdt|t|t|jdkrjtdt|dt|jt|dvrytdt||r| rtdtd d |Drtd |d td d |Drtd|d |durt||||}n%t|t rt j ||d|dd}t j ||d|dd}|d|d}t ||d} t ||d} |j } |d}|j}dd|D}|r|ddd}|ddd}n|ddd}|ddd}| rG|ddd}t||||}t||||}|rFt||||| }n#tj|||||fdd}tj|||||fdd}|rjt||||| }|st j|dg|R|| dt t j||| d}t ||||||||f}| rtj|||dntj|||d}||jd}|d d}|d d}|d|d}t j |d|dd}t j |d|dd}|d|}d||| }|||| }d|| |||| |} | rE||}!| r |!d|| || || f}!n |!d|| || f}!| | jddd|!|!jdddfS|rV| | jddd| fS| | jdddS)a%Compute Structural Similarity Index Measure. Args: preds: estimated image target: ground truth image gaussian_kernel: If true (default), a gaussian kernel is used, if false a uniform kernel is used sigma: Standard deviation of the gaussian kernel, anisotropic kernels are possible. Ignored if a uniform kernel is used kernel_size: the size of the uniform kernel, anisotropic kernels are possible. Ignored if a Gaussian kernel is used data_range: Range of the image. If ``None``, it is determined from the image (max - min) k1: Parameter of SSIM. k2: Parameter of SSIM. return_full_image: If true, the full ``ssim`` image is returned as a second argument. Mutually exclusive with ``return_contrast_sensitivity`` return_contrast_sensitivity: If true, the contrast term is returned as a second argument. The luminance term can be obtained with luminance=ssim/contrast Mutually exclusive with ``return_full_image`` rz`kernel_size` has dimension zD, but expected to be two less that target dimensionality, which is: )r*r)zMExpected `kernel_size` dimension to be 2 or 3. `kernel_size` dimensionality: zWArguments `return_full_image` and `return_contrast_sensitivity` are mutually exclusive.css$|] }|ddkp |dkVqdS)r*rNr).0xrrr ks"z_ssim_update..z8Expected `kernel_size` to have odd positive number. Got rcss|]}|dkVqdS)rNr)r+yrrrr-nsz.Expected `sigma` to have positive number. Got Nr)minmaxcSs$g|]}td|dddqS)g @g?r*r/)int)r+srrr ~s$z _ssim_update..reflect)mode)rdevice)groupsg)r0r.)ndim isinstancerrrranyr1r0tupletorchclamppowr7sizerr r Fpadr onesprodtensorcatconv3dconv2dsplitrreshapemean)"rrr!r"r#r$r%r&r'r(is_3dc1c2r7channelrgauss_kernel_sizepad_hpad_wpad_dkernel input_listoutputs output_list mu_pred_sq mu_target_sqmu_pred_target sigma_pred_sqsigma_target_sqsigma_pred_targetupperlowerssim_idx_full_imagecontrast_sensitivityrrr _ssim_update.s        $   &  $ rcelementwise_mean similarities reduction)rdsumnoneNcC t||S)aApply the specified reduction to pre-computed structural similarity. Args: similarities: per image similarities for a batch of images. reduction: a method to reduce metric score over individual batch scores - ``'elementwise_mean'``: takes the mean - ``'sum'``: takes the sum - ``'none'`` or ``None``: no reduction will be applied Returns: The reduced SSIM score r )rerfrrr _ssim_compute rjc  CsVt||\}}t||||||||| | } t| tr$| \} } t| || fS| } t| |S)aCompute Structural Similarity Index Measure. Args: preds: estimated image target: ground truth image gaussian_kernel: If true (default), a gaussian kernel is used, if false a uniform kernel is used sigma: Standard deviation of the gaussian kernel, anisotropic kernels are possible. Ignored if a uniform kernel is used kernel_size: the size of the uniform kernel, anisotropic kernels are possible. Ignored if a Gaussian kernel is used reduction: a method to reduce metric score over labels. - ``'elementwise_mean'``: takes the mean - ``'sum'``: takes the sum - ``'none'`` or ``None``: no reduction will be applied data_range: the range of the data. If None, it is determined from the data (max - min). If a tuple is provided then the range is calculated as the difference and input is clamped between the values. k1: Parameter of SSIM. k2: Parameter of SSIM. return_full_image: If true, the full ``ssim`` image is returned as a second argument. Mutually exclusive with ``return_contrast_sensitivity`` return_contrast_sensitivity: If true, the constant term is returned as a second argument. The luminance term can be obtained with luminance=ssim/contrast Mutually exclusive with ``return_full_image`` Return: Tensor with SSIM score Raises: TypeError: If ``preds`` and ``target`` don't have the same data type. ValueError: If ``preds`` and ``target`` don't have ``BxCxHxW shape``. ValueError: If the length of ``kernel_size`` or ``sigma`` is not ``2``. ValueError: If one of the elements of ``kernel_size`` is not an ``odd positive number``. ValueError: If one of the elements of ``sigma`` is not a ``positive number``. Example: >>> from torchmetrics.functional.image import structural_similarity_index_measure >>> preds = torch.rand([3, 3, 256, 256]) >>> target = preds * 0.75 >>> structural_similarity_index_measure(preds, target) tensor(0.9219) )rrcr;r=rj)rrr!r"r#rfr$r%r&r'r(similarity_pack similarityimagerrr#structural_similarity_index_measures$?  ro normalizerelusimplec CsBt||||||||dd \} } |dkrt| } t| } | | fS)NT)r(rr)rcr>rr) rrr!r"r#r$r%r&rpsimrbrrr_get_normalized_sim_and_cs's    rugǺ?g48EG?ga4?g??g9EGr?betas.c  Cs4g} |jdk} t|ts| rd|gnd|g}t|ts)| r$d|gnd|g}|ddt|ksA|ddt|krRtdt|ddt|dtd t|d d} |d| |d d krtdt|d |d d |d d | d|d| |d d krtdt|d |d d |d d | dtt|D]?} t||||||||| d \}}| |t|dkrt |d}t |d}qt|dkrt |d}t |d}qtd|| d<t | }| dkr|d d}t j||jddd }||}t j|d dS)aCompute Multi-Scale Structural Similarity Index Measure. Adapted from: https://github.com/jorge-pessoa/pytorch-msssim/blob/master/pytorch_msssim/__init__.py. Args: preds: estimated image target: ground truth image gaussian_kernel: If true, a gaussian kernel is used, if false a uniform kernel is used sigma: Standard deviation of the gaussian kernel kernel_size: size of the gaussian kernel reduction: a method to reduce metric score over labels. - ``'elementwise_mean'``: takes the mean - ``'sum'``: takes the sum - ``'none'`` or ``None``: no reduction will be applied data_range: Range of the image. If ``None``, it is determined from the image (max - min) k1: Parameter of structural similarity index measure. k2: Parameter of structural similarity index measure. betas: Exponent parameters for individual similarities and contrastive sensitives returned by different image resolutions. normalize: When MultiScaleSSIM loss is used for training, it is desirable to use normalizes to improve the training stability. This `normalize` argument is out of scope of the original implementation [1], and it is adapted from https://github.com/jorge-pessoa/pytorch-msssim instead. Raises: ValueError: If the image height or width is smaller then ``2 ** len(betas)``. ValueError: If the image height is smaller than ``(kernel_size[0] - 1) * max(1, (len(betas) - 1)) ** 2``. ValueError: If the image width is smaller than ``(kernel_size[0] - 1) * max(1, (len(betas) - 1)) ** 2``. rr)r*r9z)For a given number of `betas` parameters zH, the image height and width dimensions must be larger than or equal to rr/rz and kernel size z', the image height must be larger than z&, the image width must be larger than )rp)r*r*)r*r*r*z(length of kernel_size is neither 2 nor 3rs)r7)axis)r:r;rrArrr1rangeruappendrB avg_pool2d avg_pool3dr>stackrFr7viewrE)rrr!r"r#r$r%r&rwrpmcs_listrM _betas_div_rtrb mcs_stack mcs_weightedrrr_multiscale_ssim_updateCsX4   0          r mcs_per_imagecCri)aApply the specified reduction to pre-computed multi-scale structural similarity. Args: mcs_per_image: per image similarities for a batch of images. reduction: a method to reduce metric score over individual batch scores - ``'elementwise_mean'``: takes the mean - ``'sum'``: takes the sum - ``'none'`` or ``None``: no reduction will be applied Returns: The reduced multi-scale structural similarity r )rrfrrr_multiscale_ssim_computerkrrrc Cs|t| ts tdt| trtdd| Dstd| r%| dvr%tdt||\}}t||||||||| | } t| |S)ae Compute `MultiScaleSSIM`_, Multi-scale Structural Similarity Index Measure. This metric is a generalization of Structural Similarity Index Measure by incorporating image details at different resolution scores. Args: preds: Predictions from model of shape ``[N, C, H, W]`` target: Ground truth values of shape ``[N, C, H, W]`` gaussian_kernel: If true, a gaussian kernel is used, if false a uniform kernel is used sigma: Standard deviation of the gaussian kernel kernel_size: size of the gaussian kernel reduction: a method to reduce metric score over labels. - ``'elementwise_mean'``: takes the mean - ``'sum'``: takes the sum - ``'none'`` or ``None``: no reduction will be applied data_range: the range of the data. If None, it is determined from the data (max - min). If a tuple is provided then the range is calculated as the difference and input is clamped between the values. k1: Parameter of structural similarity index measure. k2: Parameter of structural similarity index measure. betas: Exponent parameters for individual similarities and contrastive sensitivities returned by different image resolutions. normalize: When MultiScaleSSIM loss is used for training, it is desirable to use normalizes to improve the training stability. This `normalize` argument is out of scope of the original implementation [1], and it is adapted from https://github.com/jorge-pessoa/pytorch-msssim instead. Return: Tensor with Multi-Scale SSIM score Raises: TypeError: If ``preds`` and ``target`` don't have the same data type. ValueError: If ``preds`` and ``target`` don't have ``BxCxHxW shape``. ValueError: If the length of ``kernel_size`` or ``sigma`` is not ``2``. ValueError: If one of the elements of ``kernel_size`` is not an ``odd positive number``. ValueError: If one of the elements of ``sigma`` is not a ``positive number``. Example: >>> from torch import rand >>> from torchmetrics.functional.image import multiscale_structural_similarity_index_measure >>> preds = rand([3, 3, 256, 256]) >>> target = preds * 0.75 >>> multiscale_structural_similarity_index_measure(preds, target, data_range=1.0) tensor(0.9628) References: [1] Multi-Scale Structural Similarity For Image Quality Assessment by Zhou Wang, Eero P. Simoncelli and Alan C. Bovik `MultiScaleSSIM`_ z3Argument `betas` is expected to be of a type tuple.css|]}t|tVqdS)N)r;float)r+betarrrr-szAmultiscale_structural_similarity_index_measure..z5Argument `betas` is expected to be a tuple of floats.rqzNArgument `normalize` to be expected either `None` or one of 'relu' or 'simple')r;r=rallrrr) rrr!r"r#rfr$r%r&rwrprrrr.multiscale_structural_similarity_index_measures E  r)TrrNrr FF)rd) TrrrdNrr FF)TrrNrr N)TrrNrr rvN) TrrrdNrr rvrr)!collections.abcrtypingrrrr>rtorch.nnrrBtyping_extensionsr#torchmetrics.functional.image.utilsr r r torchmetrics.utilities.checksr "torchmetrics.utilities.distributedrr=rboolrr2rcrjrorurrrrrrrsr              X     k