3j@SSKJr SSKrSSKJs Jr SSKJr SSKJrJ r J r SSK J r J r SSKJr S\R S\\S \R 4S jrSS \R S \R S \S\S\S\S \R 4Sjjr"SS\R,5rg))ListN)nn) KORNIA_CHECKKORNIA_CHECK_IS_TENSORKORNIA_CHECK_SHAPE)filter3dget_gaussian_kernel3d)_compute_paddingimgcropping_shapereturnc l[R"UUS*US*US*US*US*US*45$)z.Crop out the part of "valid" convolution area.r)Fpad)r r s O/home/wildlama/miniconda3/lib/python3.13/site-packages/kornia/metrics/ssim3d.py_croprs] 55 A   A   A   A   A   A      img1img2 window_sizemax_valepspaddingct[U5 [U5 [U/SQ5 [U/SQ5 [URUR:HSURSUR35 [ U[ 5(d[ S[U535e[X"U4S5nSU-S-nSU-S-n[X5n [X5n /n US :Xa6URS S upn[XU/5n [X5n [X5n OUS :XaU S-nU S-nX-n[US-U5n[US-U5n[X-U5nUS :Xa%[UU 5n[UU 5n[UU 5nOUS :XaUU- nUU- nUU- nS U-U-S U-U--nUU-U-UU-U--nUUU-- $)aCompute the Structural Similarity (SSIM) index map between two images. Measures the (SSIM) index between each element in the input `x` and target `y`. The index can be described as: .. math:: \text{SSIM}(x, y) = \frac{(2\mu_x\mu_y+c_1)(2\sigma_{xy}+c_2)} {(\mu_x^2+\mu_y^2+c_1)(\sigma_x^2+\sigma_y^2+c_2)} torch.where: - :math:`c_1=(k_1 L)^2` and :math:`c_2=(k_2 L)^2` are two variables to stabilize the division with weak denominator. - :math:`L` is the dynamic range of the pixel-values (typically this is :math:`2^{\#\text{bits per pixel}}-1`). Args: img1: the first input image with shape :math:`(B, C, D, H, W)`. img2: the second input image with shape :math:`(B, C, D, H, W)`. window_size: the size of the gaussian kernel to smooth the images. max_val: the dynamic range of the images. eps: Small value for numerically stability when dividing. padding: ``'same'`` | ``'valid'``. Whether to only use the "valid" convolution area to compute SSIM to match the MATLAB implementation of original SSIM paper. Returns: The ssim index map with shape :math:`(B, C, D, H, W)`. Examples: >>> input1 = torch.rand(1, 4, 5, 5, 5) >>> input2 = torch.rand(1, 4, 5, 5, 5) >>> ssim_map = ssim3d(input1, input2, 5) # 1x4x5x5x5 )BCDHWz,img1 and img2 shapes must be the same. Got: z and z'Input max_val type is not a float. Got )?r%r%g{Gz?rgQ?validNsameg@) rrrshape isinstancefloat TypeErrortyper rr r)rrrrrrkernelC1C2mu1mu2r depthheightwidthmu1_sqmu2_sqmu1_mu2 mu_img1_sq mu_img2_sq mu_img1_img2 sigma1_sq sigma2_sqsigma12numdens rssim3drA,sV4 4 t67t67tzz)-YZ^ZdZdYeejkokukujv+wx gu % %A$w-QRR1+K1XZijFA%BA%B!.C .C "N'%||BC0u)5%*@AC(C( F  !VF !VFiG$'6*J$'6*JDK0L':~6 :~6 \>: F  V#IV#IW$Gw+g 0BCC&2-)i2G"2LMC #) rc ^\rSrSrSrSS\S\S\S\SS4 U4S jjjrS \ RS \ RS\ R4S jr S r U=r $)SSIM3DaCreate a module that computes the Structural Similarity (SSIM) index between two 3D images. Measures the (SSIM) index between each element in the input `x` and target `y`. The index can be described as: .. math:: \text{SSIM}(x, y) = \frac{(2\mu_x\mu_y+c_1)(2\sigma_{xy}+c_2)} {(\mu_x^2+\mu_y^2+c_1)(\sigma_x^2+\sigma_y^2+c_2)} torch.where: - :math:`c_1=(k_1 L)^2` and :math:`c_2=(k_2 L)^2` are two variables to stabilize the division with weak denominator. - :math:`L` is the dynamic range of the pixel-values (typically this is :math:`2^{\#\text{bits per pixel}}-1`). Args: window_size: the size of the gaussian kernel to smooth the images. max_val: the dynamic range of the images. eps: Small value for numerically stability when dividing. padding: ``'same'`` | ``'valid'``. Whether to only use the "valid" convolution area to compute SSIM to match the MATLAB implementation of original SSIM paper. Shape: - Input: :math:`(B, C, D, H, W)`. - Target :math:`(B, C, D, H, W)`. - Output: :math:`(B, C, D, H, W)`. Examples: >>> input1 = torch.rand(1, 4, 5, 5, 5) >>> input2 = torch.rand(1, 4, 5, 5, 5) >>> ssim = SSIM3D(5) >>> ssim_map = ssim(input1, input2) # 1x4x5x5x5 rrrrr NcR>[TU]5 XlX lX0lX@lgN)super__init__rrrr)selfrrrr __class__s rrHSSIM3D.__init__s#  +%  rrrcp[XURURURUR5$rF)rArrrr)rIrrs rforwardSSIM3D.forwards'd$"2"2DLL$((DLLYYr)rrrrg?g-q=r()__name__ __module__ __qualname____firstlineno____doc__intr+strrHtorchTensorrM__static_attributes__ __classcell__)rJs@rrCrCsh#JC%E\_mqZELLZ ZZZrrCrO)typingrrWtorch.nn.functionalr functionalrkornia.core.checkrrrkornia.filtersrr kornia.filters.filterr rXrUrr+rVrAModulerCrrrcs$ VV:2 u|| T#Y 5<< & ` ,,` ,,`` `  `  ` \\`F.ZRYY.Zr