Skip to content

[Inference.Core] HED Soft-Edge Lines

Documentation

  • Class name: Inference_Core_HEDPreprocessor
  • Category: ControlNet Preprocessors/Line Extractors
  • Output node: False

The HEDPreprocessor node is designed for extracting soft-edge lines from images using the HED (Holistically-Nested Edge Detection) model. It preprocesses images to enhance edge features, facilitating downstream tasks that require detailed edge or line information.

Input types

Required

  • image
    • The input image to be processed for edge detection. The resolution parameter allows specifying the desired output resolution, affecting the level of detail in the detected edges.
    • Comfy dtype: IMAGE
    • Python dtype: torch.Tensor

Optional

  • safe
    • A mode toggle that enables or disables safety checks during edge detection, potentially affecting performance and output quality.
    • Comfy dtype: COMBO[STRING]
    • Python dtype: str
  • resolution
    • Specifies the resolution of the output image, influencing the granularity of detected edges. Higher resolutions yield more detailed edge information.
    • Comfy dtype: INT
    • Python dtype: int

Output types

  • image
    • Comfy dtype: IMAGE
    • The processed image with enhanced edge details, suitable for use in various image analysis and manipulation tasks.
    • Python dtype: torch.Tensor

Usage tips

  • Infra type: GPU
  • Common nodes: unknown

Source code

class HED_Preprocessor:
    @classmethod
    def INPUT_TYPES(s):
        return create_node_input_types(
            safe=(["enable", "disable"], {"default": "enable"})
        )

    RETURN_TYPES = ("IMAGE",)
    FUNCTION = "execute"

    CATEGORY = "ControlNet Preprocessors/Line Extractors"

    def execute(self, image, resolution=512, **kwargs):
        from controlnet_aux.hed import HEDdetector

        model = HEDdetector.from_pretrained().to(model_management.get_torch_device())
        out = common_annotator_call(model, image, resolution=resolution, safe = kwargs["safe"] == "enable")
        del model
        return (out, )