3j$#SSKJr SSKJrJrJr SSKrSSKJr SSKJ r SSK J r SSK J r SSKJr SS KJr SS KJr SS KJr SS KJrJr S S/r"SS \\5r"SS5rg)) annotations)AnyOptionalUnionN)nn)grayscale_to_rgb)PILImage)onnx)ONNXExportMixin) Normalize) ModelBase)DexiNed)ResizePostProcessorResizePreProcessor EdgeDetectorEdgeDetectorBuilderc^\rSrSr%SrSrS\S'S SU4Sjjjr\SSj5r \ R"5SSj5r SSS jjr SSS jjrSSS jjrS rU=r$)r%aEdgeDetector is a module that wraps an edge detection model. This is a high-level API that wraps edge detection models like :py:class:`kornia.models.DexiNed`. Args: model: The edge detection model. pre_processor: Pre-processing module (e.g., ResizePreProcessor). post_processor: Post-processing module (e.g., ResizePostProcessor). name: Optional name for the detector. Example: >>> from kornia.models.dexined import DexiNed >>> from kornia.models.processors import ResizePreProcessor, ResizePostProcessor >>> model = DexiNed(pretrained=True) >>> detector = EdgeDetector(model, ResizePreProcessor(352, 352), ResizePostProcessor()) >>> img = torch.rand(1, 3, 320, 320) >>> out = detector(img) edge_detectionstrnamecx>[TU]5 UR5UlX lX0lUbX@lgg)zInitialize EdgeDetector. Args: model: The edge detection model. pre_processor: Pre-processing module (e.g., ResizePreProcessor). post_processor: Post-processing module (e.g., ResizePostProcessor). name: Optional name for the detector. N)super__init__evalmodel pre_processorpost_processorr)selfrrrr __class__s W/home/wildlama/miniconda3/lib/python3.13/site-packages/kornia/contrib/edge_detection.pyrEdgeDetector.__init__<s9 ZZ\ *,  I c[S5e)a%Build EdgeDetector from config. This is a placeholder to satisfy the abstract method requirement. Use EdgeDetectorBuilder.build() or instantiate EdgeDetector directly. Args: config: Configuration object (not used, kept for interface compatibility). Returns: EdgeDetector instance. ztEdgeDetector.from_config() is not implemented. Use EdgeDetectorBuilder.build() or instantiate EdgeDetector directly.)NotImplementedError)configs r! from_configEdgeDetector.from_configSs" T  r#clURU5upURU5nURX25$)zForward pass of the edge detection model. Args: images: If list of RGB images. Each image is a torch.Tensor with shape :math:`(3, H, W)`. If torch.Tensor, a torch.Tensor with shape :math:`(B, 3, H, W)`. Returns: output torch.Tensor. )rrr)rimages image_sizes out_imagess r!forwardEdgeDetector.forwardfs6#008ZZ' "":;;r#cUcURU5n/nUH nUR[U5S5 M" URXC[ U[ R 5S9$)zDraw the edge detection results. Args: images: input torch.Tensor. edge_maps: detected edges. output_type: type of the output. Returns: output torch.Tensor. r)is_batch)r-appendr_tensor_to_type isinstancetorchTensor)rr* edge_maps output_typeoutputedge_maps r! visualizeEdgeDetector.visualizevsd"   V,I!H MM*84Q7 8"##F*VUZUaUaBb#ccr#chURXU5nURXSS9 URXSSS9 g)zSave the edge detection results. Args: images: input torch.Tensor. edge_maps: detected edges. output_type: type of the output. directory: where to save outputs. Returns: output torch.Tensor. _src)suffix_edgeN)r: _save_outputs)rr*r6 directoryr7outputss r!saveEdgeDetector.saves:&..K@ 6V< 7g>r#c UcSURSUS3n[R"UU4SSU=(d SU=(d S/SSU=(d SU=(d S/SSU=(d SU=(d S/U(aUO URUUS.UD6$) aExport the current edge detection model to an ONNX model file. Args: onnx_name: The name of the output ONNX file. If not provided, a default name in the format "Kornia-.onnx" will be used. image_size: The size to which input images will be resized during preprocessing. If None, image_size will be dynamic. For DexiNed, recommended scale is 352. include_pre_and_post_processor: Whether to include the pre-processor and post-processor in the exported model. save: If to save the model or load it. additional_metadata: Additional metadata to add to the ONNX model. kwargs: Additional arguments to convert to onnx. kornia__z.onnx`) input_shape output_shape pseudo_shaperrCadditional_metadata)rr to_onnxr)r onnx_name image_sizeinclude_pre_and_post_processorrCrOkwargss r!rPEdgeDetector.to_onnxs6  !$))Aj\?I&&    Q 0b*2BCa!1r:3CDQ 1c:3DE8$djj 3    r#)rrrr)N) rtorch.nn.ModulerrVrrVr Optional[str]returnNone)r&rrXr)r*'Union[torch.Tensor, list[torch.Tensor]]rXrZ)Nr4)r*rZr61Optional[Union[torch.Tensor, list[torch.Tensor]]]r7rrXz:Union[torch.Tensor, list[torch.Tensor], list[Image.Image]])NNr4) r*rZr6r[rArWr7rrXrY)NrKTTN)rQrWrRz Optional[int]rSboolrCr\rOzOptional[list[tuple[str, str]]]rTrrXzonnx.ModelProto)__name__ __module__ __qualname____firstlineno____doc__r__annotations__r staticmethodr'r4inference_moder-r:rCrP__static_attributes__ __classcell__)r s@r!rr%sb(!D# # '(     .  $  < <$HL" d7dEd d D d8HL#'" ?7?E?! ?  ?  ?2$($'/3?C (  ( "( )- (  ( = ( (  ( ( r#c0\rSrSrSr\SSSjj5rSrg)ra'EdgeDetectorBuilder is a class that builds an edge detection model. This is a high-level API that builds edge detection models like :py:class:`kornia.models.DexiNed` and wraps them with :py:class:`EdgeDetector`. Note: To use this model, load image tensors and call ``model.save(images)``. c\UR5S:Xam[[R"/SQ/5[R"S/S-/5S9n[R "U[ US9[R"55nO[SUS35e[U[X"5[5SS 9$) aBuild an edge detection model. Args: model_name: Name of the model to build. Currently only "dexined" is supported. pretrained: If True, loads pretrained weights. image_size: Size to which input images will be resized during preprocessing. Returns: EdgeDetector instance configured with the specified model. Example: >>> detector = EdgeDetectorBuilder.build(pretrained=True, image_size=352) >>> img = torch.rand(1, 3, 320, 320) >>> out = detector(img) dexined)g ףp= ?gv/?gCl?gp?rI)meanstd) pretrainedzModel z) not found. Please choose from 'dexined'.)r) lowerr r4tensorr SequentialrSigmoid ValueErrorrrr) model_namermrRnormrs r!buildEdgeDetectorBuilder.builds$     *%,,0E/F"GU\\\g[hkl[lZmMnoDMM$:(F UEvj\1Z[\ \  z 6  !   r#N)rjTrK)rsrrmr\rRintrXr)r]r^r_r`rarcrurerwr#r!rrs  r#) __future__rtypingrrrr4rkornia.color.grayrkornia.core.externalr Imager kornia.core.mixin.onnxr kornia.enhance.normalizer kornia.models.baser kornia.models.dexinedrkornia.models.processorsrr__all__rrrwr#r!rsT$#'' .2%2.()L 0 1i 9oi X( ( r#