85 lines
3.8 KiB
Markdown
85 lines
3.8 KiB
Markdown
# Documentation
|
|
- Class name: JagsCombineMasks
|
|
- Category: Jags_vector/CLIPSEG
|
|
- Output node: False
|
|
- Repo Ref: https://github.com/jags111/ComfyUI_Jags_VectorMagic
|
|
|
|
该节点将多个掩码张量协同整合为统一的表示形式，便于在给定的图像上下文中集成分割区域。它旨在简化将二进制或分类数据合并为一致视觉输出的过程，提高基于掩码的图像处理任务的实用性。
|
|
|
|
# Input types
|
|
## Required
|
|
- input_image
|
|
- 输入图像作为合并和可视化掩码的基础层，其作用至关重要，因为它提供了准确叠加和整合掩码所需的空间上下文。
|
|
- Comfy dtype: IMAGE
|
|
- Python dtype: torch.Tensor
|
|
- mask_1
|
|
- 主掩码在融合过程中是一个关键组件，定义了初始分割，为后续叠加其他掩码奠定了基础。
|
|
- Comfy dtype: MASK
|
|
- Python dtype: torch.Tensor
|
|
- mask_2
|
|
- 次级掩码通过引入额外的分割层来完善合并掩码，提高了最终输出的粒度和特异性。
|
|
- Comfy dtype: MASK
|
|
- Python dtype: torch.Tensor
|
|
## Optional
|
|
- mask_3
|
|
- 可选的第三级掩码为掩码组合提供了额外的维度，当需要时允许更复杂和细致的分割。
|
|
- Comfy dtype: MASK
|
|
- Python dtype: Optional[torch.Tensor]
|
|
|
|
# Output types
|
|
- Combined Mask
|
|
- 生成的合并掩码包含了输入掩码的集体分割信息，作为图像中分割元素的综合表示。
|
|
- Comfy dtype: MASK
|
|
- Python dtype: torch.Tensor
|
|
- Heatmap Mask
|
|
- 热图掩码通过颜色叠加可视化合并分割，提供了直观且易于解释的分割结果描述。
|
|
- Comfy dtype: IMAGE
|
|
- Python dtype: torch.Tensor
|
|
- BW Mask
|
|
- 二值掩码提供了一个清晰、高对比度的分割表示，以清晰和精确的方式突出显示划分的区域。
|
|
- Comfy dtype: IMAGE
|
|
- Python dtype: torch.Tensor
|
|
|
|
# Usage tips
|
|
- Infra type: CPU
|
|
|
|
# Source code
|
|
```
|
|
class JagsCombineMasks:
|
|
|
|
def __init__(self):
|
|
pass
|
|
|
|
@classmethod
|
|
def INPUT_TYPES(s):
|
|
return {'required': {'input_image': ('IMAGE',), 'mask_1': ('MASK',), 'mask_2': ('MASK',)}, 'optional': {'mask_3': ('MASK',)}}
|
|
CATEGORY = 'Jags_vector/CLIPSEG'
|
|
RETURN_TYPES = ('MASK', 'IMAGE', 'IMAGE')
|
|
RETURN_NAMES = ('Combined Mask', 'Heatmap Mask', 'BW Mask')
|
|
FUNCTION = 'combine_masks'
|
|
|
|
def combine_masks(self, input_image: torch.Tensor, mask_1: torch.Tensor, mask_2: torch.Tensor, mask_3: Optional[torch.Tensor]=None) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
|
"""A method that combines two or three masks into one mask. Takes in tensors and returns the mask as a tensor, as well as the heatmap and binary mask as tensors."""
|
|
if mask_1 is not None:
|
|
mask_1 = mask_1.squeeze()
|
|
if mask_2 is not None:
|
|
mask_2 = mask_2.squeeze()
|
|
if mask_3 is not None:
|
|
mask_3 = mask_3.squeeze()
|
|
combined_mask = mask_1 + mask_2 + mask_3 if mask_3 is not None else mask_1 + mask_2
|
|
image_np = tensor_to_numpy(input_image)
|
|
heatmap = apply_colormap(combined_mask, cm.viridis)
|
|
binary_mask = apply_colormap(combined_mask, cm.Greys_r)
|
|
dimensions = (image_np.shape[1], image_np.shape[0])
|
|
print('heatmap', heatmap)
|
|
if dimensions is None or dimensions[0] == 0 or dimensions[1] == 0:
|
|
raise ValueError('Invalid dimensions')
|
|
heatmap_resized = resize_image(heatmap, dimensions)
|
|
binary_mask_resized = resize_image(binary_mask, dimensions)
|
|
(alpha_heatmap, alpha_binary) = (0.5, 1)
|
|
overlay_heatmap = overlay_image(image_np, heatmap_resized, alpha_heatmap)
|
|
overlay_binary = overlay_image(image_np, binary_mask_resized, alpha_binary)
|
|
image_out_heatmap = numpy_to_tensor(overlay_heatmap)
|
|
image_out_binary = numpy_to_tensor(overlay_binary)
|
|
return (combined_mask, image_out_heatmap, image_out_binary)
|
|
``` |