Inpaint

Documentation

• Class name: Inpaint
• Category: Bmad/CV/C.Photography
• Output node: False

The Inpaint node is designed to reconstruct the missing or damaged parts of images using a specified inpainting technique. It leverages computer vision algorithms to fill in gaps or remove unwanted objects from images, based on the surrounding pixel information.

Input types

Required

• img
  • The image to be inpainted. It serves as the primary input where missing parts or imperfections are identified and subsequently reconstructed.
  • Comfy dtype: IMAGE
  • Python dtype: torch.Tensor
• mask
  • A mask image indicating the areas to be inpainted. Pixels marked in the mask are the targets for inpainting, guiding the algorithm on where to apply the reconstruction process.
  • Comfy dtype: IMAGE
  • Python dtype: torch.Tensor
• radius
  • Defines the neighborhood size around each point to consider for inpainting. A larger radius uses more surrounding pixels for reconstruction, potentially leading to smoother results.
  • Comfy dtype: INT
  • Python dtype: int
• flag
  • Specifies the inpainting technique to use. Different methods can produce varying results based on the algorithm's approach to filling in missing data.
  • Comfy dtype: COMBO[STRING]
  • Python dtype: str

Output types

• image
  • Comfy dtype: IMAGE
  • The inpainted image, where the specified areas have been reconstructed using the chosen inpainting technique.
  • Python dtype: torch.Tensor

Usage tips

• Infra type: CPU
• Common nodes: unknown

Source code

class Inpaint:
    inpaint_method_map = {
        "TELEA": cv.INPAINT_TELEA,
        "NS": cv.INPAINT_NS,
    }
    inpaint_methods = list(inpaint_method_map.keys())

    @classmethod
    def INPUT_TYPES(cls):
        return {
            "required": {
                "img": ("IMAGE",),
                "mask": ("IMAGE",),
                "radius": ("INT", {"default": 3, "min": 0, "step": 1}),
                "flag": (cls.inpaint_methods, {"default": cls.inpaint_methods[0]}),
            },
        }

    RETURN_TYPES = ("IMAGE",)
    FUNCTION = "paint"
    CATEGORY = f"{cv_category_path}/C.Photography"

    def paint(self, img, mask, radius, flag):
        img = tensor2opencv(img)
        mask = tensor2opencv(mask, 1)
        dst = cv.inpaint(img, mask, radius, self.inpaint_method_map[flag])
        result = opencv2tensor(dst)
        return (result,)