LayerMask: BiRefNetUltra

Documentation

• Class name: LayerMask: BiRefNetUltra
• Category: 😺dzNodes/LayerMask
• Output node: False

The BiRefNetUltra node is designed for advanced image processing, specifically focusing on background removal and refinement of layer masks. It leverages the BiRefNet architecture to achieve high-quality segmentation and mask generation, catering to applications requiring precise foreground-background separation.

Input types

Required

• image
  • The primary image input for processing, which the node uses to generate and refine masks based on the background and foreground segmentation.
  • Comfy dtype: IMAGE
  • Python dtype: torch.Tensor
• detail_method
  • Specifies the method used for enhancing the detail of the mask, including options like 'VITMatte', 'VITMatte(local)', 'PyMatting', and 'GuidedFilter', to adjust the mask's precision and quality.
  • Comfy dtype: COMBO[STRING]
  • Python dtype: str
• detail_erode
  • Determines the erosion level applied to the mask, affecting the mask's edge by reducing its size to remove small, unwanted foreground elements.
  • Comfy dtype: INT
  • Python dtype: int
• detail_dilate
  • Determines the dilation level applied to the mask, affecting the mask's edge by expanding its size to include additional foreground elements that might have been missed.
  • Comfy dtype: INT
  • Python dtype: int
• black_point
  • Sets the black point for histogram remapping, influencing the darkest areas of the mask to refine shadow details and contrast.
  • Comfy dtype: FLOAT
  • Python dtype: float
• white_point
  • Sets the white point for histogram remapping, influencing the brightest areas of the mask to enhance highlights and details.
  • Comfy dtype: FLOAT
  • Python dtype: float
• process_detail
  • A boolean flag that indicates whether to apply detailed processing to the mask, affecting the overall quality and precision of the mask refinement.
  • Comfy dtype: BOOLEAN
  • Python dtype: bool
• device
  • Specifies the computing device ('cuda' or 'cpu') used for processing, affecting performance and efficiency.
  • Comfy dtype: COMBO[STRING]
  • Python dtype: str
• max_megapixels
  • Limits the maximum size of the image in megapixels to optimize processing time and resource usage, affecting the resolution of the processed image and mask.
  • Comfy dtype: FLOAT
  • Python dtype: float

Optional

Output types

• image
  • Comfy dtype: IMAGE
  • The processed image with refined foreground-background separation, suitable for further use or visualization.
  • Python dtype: torch.Tensor
• mask
  • Comfy dtype: MASK
  • The generated mask that highlights the foreground by separating it from the background, ready for use in compositing or further processing.
  • Python dtype: torch.Tensor

Usage tips

• Infra type: GPU
• Common nodes: unknown

Source code

class BiRefNetUltra:

    @classmethod
    def INPUT_TYPES(cls):

        method_list = ['VITMatte', 'VITMatte(local)', 'PyMatting', 'GuidedFilter', ]
        device_list = ['cuda','cpu']
        return {
            "required": {
                "image": ("IMAGE",),
                "detail_method": (method_list,),
                "detail_erode": ("INT", {"default": 6, "min": 1, "max": 255, "step": 1}),
                "detail_dilate": ("INT", {"default": 6, "min": 1, "max": 255, "step": 1}),
                "black_point": ("FLOAT", {"default": 0.01, "min": 0.01, "max": 0.98, "step": 0.01, "display": "slider"}),
                "white_point": ("FLOAT", {"default": 0.99, "min": 0.02, "max": 0.99, "step": 0.01, "display": "slider"}),
                "process_detail": ("BOOLEAN", {"default": True}),
                "device": (device_list,),
                "max_megapixels": ("FLOAT", {"default": 2.0, "min": 1, "max": 999, "step": 0.1}),
            },
            "optional": {
            }
        }

    RETURN_TYPES = ("IMAGE", "MASK", )
    RETURN_NAMES = ("image", "mask", )
    FUNCTION = "birefnet_ultra"
    CATEGORY = '😺dzNodes/LayerMask'

    def birefnet_ultra(self, image, detail_method, detail_erode, detail_dilate,
                       black_point, white_point, process_detail, device, max_megapixels):
        ret_images = []
        ret_masks = []

        if detail_method == 'VITMatte(local)':
            local_files_only = True
        else:
            local_files_only = False

        from .birefnet_func import BiRefNetRemoveBackground
        birefnetrmbg = BiRefNetRemoveBackground()

        for i in image:
            i = torch.unsqueeze(i, 0)
            orig_image = tensor2pil(i).convert('RGB')

            _mask = birefnetrmbg.generate_mask(orig_image)
            _mask = image2mask(_mask)

            detail_range = detail_erode + detail_dilate

            if process_detail:
                if detail_method == 'GuidedFilter':
                    _mask = guided_filter_alpha(i, _mask, detail_range // 6 + 1)
                    _mask = tensor2pil(histogram_remap(_mask, black_point, white_point))
                elif detail_method == 'PyMatting':
                    _mask = tensor2pil(mask_edge_detail(i, _mask, detail_range // 8 + 1, black_point, white_point))
                else:
                    _trimap = generate_VITMatte_trimap(_mask, detail_erode, detail_dilate)
                    _mask = generate_VITMatte(orig_image, _trimap, local_files_only=local_files_only, device=device, max_megapixels=max_megapixels)
                    _mask = tensor2pil(histogram_remap(pil2tensor(_mask), black_point, white_point))
            else:
                _mask = tensor2pil(_mask)

            ret_image = RGB2RGBA(orig_image, _mask.convert('L'))
            ret_images.append(pil2tensor(ret_image))
            ret_masks.append(image2mask(_mask))

        log(f"{NODE_NAME} Processed {len(ret_masks)} image(s).", message_type='finish')
        return (torch.cat(ret_images, dim=0), torch.cat(ret_masks, dim=0),)