3j `SSKrSSKJs Jr SS\R S\S\R 4Sjjrg)Nimagenum_iterationsreturnc[U[R5(d[S[ U535e[U[ 5(aUS:a [S5e[ UR5S:dURSS:wa[SUR35eURSSup#URS SX#5nUS:HnURun n[R"SXb-U-S-URURS 9RS SX#45nS X)'[U5H0n[R"USSSS 9n[R "X5nM2 UR#U5$) ajCompute the Connected-component labelling (CCL) algorithm. .. image:: https://github.com/kornia/data/raw/main/cells_segmented.png The implementation is an adaptation of the following repository: https://gist.github.com/efirdc/5d8bd66859e574c683a504a4690ae8bc .. warning:: This is an experimental API subject to changes and optimization improvements. .. note:: See a working example `here `__. Args: image: the binarized input image with shape :math:`(*, 1, H, W)`. The image must be in floating point with range [0, 1]. num_iterations: the number of iterations to make the algorithm to converge. Return: The labels image with the same shape of the input image. Example: >>> img = torch.rand(2, 1, 4, 5) >>> img_labels = connected_components(img, num_iterations=100) z,Input imagetype is not a torch.Tensor. Got: z0Input num_iterations must be a positive integer.z*Input image shape must be (*,1,H,W). Got: N)devicedtyper) kernel_sizestridepadding) isinstancetorchTensor TypeErrortypeintlenshape ValueErrorviewaranger r rangeF max_pool2dmulview_as) rrHW image_viewmaskB_outs ]/home/wildlama/miniconda3/lib/python3.13/site-packages/kornia/contrib/connected_components.pyconnected_componentsr)sM8 eU\\ * *FtE{mTUU nc * *nq.@JKK 5;;!u{{2!3Eekk]STT ;;rs DAB1(J ?D!!JAq!Q ,,q!%!)a- EKK P U UWY[\^_Vc dCCJ > "ll3AaCii"# ;;u )d)rtorch.nn.functionalnn functionalrrrr)r*r(r0s2$ 4 4c4ELL4r*