v1.1.0: Optimize dependency management and improve documentation

2025-06-09 18:10:23 +08:00
parent a35c8b7673
commit 9da405414a
5 changed files with 98 additions and 92 deletions
@@ -1,92 +1,87 @@
 # ComfyUI-RemoveBackgroundSuite
-![image](https://github.com/user-attachments/assets/75fb08b6-184d-46e1-a4e6-3294dd94d66d)

-> 基于 ComfyUI 的抠图套件，支持多种抠图模型和细节处理方式。
+A matting toolkit based on ComfyUI, supporting multiple matting models and detail processing methods.

-## 版本信息
+## Features

- 当前版本：v0.2
- 更新日期：2024-06-XX
+- **Multiple Models**: Support for various matting models including BiRefNet, RMBG, and more
+- **Detail Processing**: Advanced mask processing capabilities for fine-tuning results
+- **User-Friendly**: Simple and intuitive interface within ComfyUI
+- **High Performance**: Optimized for both quality and speed

-## 更新日志
+## Installation

-### v0.2 (2024-06-XX)
- BiRefNetUltraV3_RBS 节点合并模型加载功能，直接选择模型版本，无需单独加载节点。
- BiRefNetUltraV3_RBS、TransparentBackgroundUltra_RBS 节点的 max_megapixels 参数支持动态调整分辨率，自动缩放为32的倍数，防止patch分割报错。
- 新增 Mask Process Details (RBS) 节点，支持VITMatte/GuidedFilter等多种细节处理方法，参数可调。
- 节点命名与界面优化，部分节点重命名。
- 细节处理参数范围优化。
-
-### v1.2 (2024-06-XX)
- 节点界面极简化，移除所有细节处理相关参数和功能。
- TransparentBackgroundUltra_RBS 去除 WIP 标志。
-
-### v1.1 (2024-03-21)
- 优化 BiRefNet 模型加载逻辑，支持 dynamic、HR、HR-matting 模型
- 优化 VITMatte 模型加载和推理流程
- 改进文档结构和说明
-
-### v1.0 (2024-03-20)
- 初始版本发布
- 支持 BiRefNet 和 TransparentBackground 两种抠图模型
-
-## 节点说明
-
-### 1. BiRefNet Ultra V3 (RBS)
- **功能**：使用 BiRefNet Ultra V3 进行高质量背景移除，支持 dynamic/HR/HR-matting 等多种模型。
- **参数**：
-  - `image`：输入图片（支持批量）。
-  - `version`：选择模型版本（BiRefNet-General、RMBG-2.0、BiRefNet_dynamic等）。
-  - `device`、`max_megapixels`：详见节点界面。max_megapixels 会自动动态缩放分辨率为32的倍数。
- **输出**：
-  - `image`：去背景后的 RGBA 图片
-  - `mask`：前景掩码
-
-### 2. Transparent Background Ultra (RBS)
- **功能**：将图片背景转换为透明。
- **参数**：
-  - `image`：输入图片。
-  - `model`：选择本地模型。
-  - `device`、`max_megapixels`：详见节点界面。max_megapixels 会自动动态缩放分辨率为32的倍数。
- **输出**：
-  - `image`：透明背景图片
-  - `mask`：前景掩码
-
-### 3. Mask Process Details (RBS)
- **功能**：对输入图像和mask进行细节优化，支持VITMatte、GuidedFilter等多种方法。
- **参数**：
-  - `image`：输入图片。
-  - `mask`：输入掩码。
-  - `detail_method`：细节处理方法（VITMatte、VITMatte(local)、GuidedFilter）。
-  - `detail_erode`、`detail_dilate`、`black_point`、`white_point`、`device`、`max_megapixels`：详见节点界面。
- **输出**：
-  - `image`：优化后的图片
-  - `mask`：优化后的掩码
-
-## 典型用法
-1. 用 `BiRefNet Ultra V3 (RBS)` 进行背景移除。
-2. 可选：用 `Mask Process Details (RBS)` 进一步优化边缘细节。
-3. 可选：用 `Transparent Background Ultra (RBS)` 处理透明背景。
-
-## 注意事项
- 请将模型文件放在 `ComfyUI/models/transparent-background/` 目录下，或使用新版节点自动下载。
- 推荐使用 CUDA 设备以获得更快推理速度。
- 细节处理方法对边缘质量有显著影响，可根据实际需求调整。
- 插件所有节点均归类于 `RemoveBackgroundSuite`，便于统一管理。
-
-## 依赖安装
+1. Navigate to your ComfyUI's `custom_nodes` directory
+2. Clone this repository:
 ```bash
+git clone https://github.com/whmc76/ComfyUI-RemoveBackgroundSuite.git
+```
+3. Install dependencies:
+```bash
+cd ComfyUI-RemoveBackgroundSuite
 pip install -r requirements.txt
 ```

-## 常见问题
- **模型下载失败**：请检查网络连接或手动下载模型放入指定目录。
- **推理慢/显存不足**：可适当降低 `max_megapixels` 或切换到 CPU。
- **节点不显示**：请确认插件已放入 `custom_nodes` 目录并重启 ComfyUI。
+## Usage

---
+1. Start ComfyUI
+2. The new nodes will appear in the node menu under the "RBS" category
+3. Connect the nodes as needed in your workflow

-## 致谢
-本插件大量借鉴和参考了 [ComfyUI_LayerStyle_Advance](https://github.com/chflame163/ComfyUI_LayerStyle_Advance) 项目的设计与实现，特别感谢原作者 chflame163 的开源贡献！
+## Models

-如有更多问题请参考原项目文档或在 Issues 区反馈。
+The following models are supported:
+
+- BiRefNet-General
+- BiRefNet_dynamic
+- BiRefNet_HR
+- BiRefNet_HR-matting
+- RMBG-2.0
+
+## Nodes
+
+### BiRefNetUltraV3_RBS
+- **Input**: Image
+- **Output**: Mask
+- **Parameters**:
+  - Model Version: Select from available BiRefNet models
+  - Max Megapixels: Maximum image size for processing
+
+### Transparent Background Ultra (RBS)
+- **Input**: Image
+- **Output**: Transparent Image
+- **Parameters**:
+  - Model Version: Select from available models
+  - Max Megapixels: Maximum image size for processing
+
+### Mask Process Details (RBS)
+- **Input**: Mask
+- **Output**: Processed Mask
+- **Parameters**:
+  - Blur Radius: Gaussian blur radius for mask smoothing
+  - Feather Radius: Edge feathering radius
+  - Contrast: Mask contrast adjustment
+  - Brightness: Mask brightness adjustment
+
+## Changelog
+
+### v1.1.0
+- Optimized dependency management
+- Removed version constraints for better compatibility
+- Removed unused dependencies
+- Improved code organization
+
+### v1.0.0
+- Initial release with core functionality
+- Support for multiple matting models
+- Basic mask processing capabilities
+
+## License
+
+This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
+
+## Acknowledgments
+
+- [ComfyUI](https://github.com/comfyanonymous/ComfyUI) for the amazing framework
+- [BiRefNet](https://github.com/ZhengPeng7/BiRefNet) for the matting models
+- [RMBG](https://github.com/briaai/RMBG-2.0) for the background removal model
@@ -2,4 +2,4 @@ from .nodes import NODE_CLASS_MAPPINGS, NODE_DISPLAY_NAME_MAPPINGS

 __all__ = ['NODE_CLASS_MAPPINGS', 'NODE_DISPLAY_NAME_MAPPINGS']

-__version__ = 'v0.2' 
+__version__ = 'v1.1.0' 
@@ -240,7 +240,7 @@ class ProcessDetails_RBS:
            "required": {
                "image": ("IMAGE",),
                "mask": ("MASK",),
-                "detail_method": (["VITMatte", "VITMatte(local)", "GuidedFilter"], {"default": "VITMatte"}),
+                "detail_method": (["VITMatte", "VITMatte(local)", "PyMatting", "GuidedFilter"], {"default": "VITMatte"}),
                "detail_erode": ("INT", {"default": 4, "min": 1, "max": 100, "step": 1}),
                "detail_dilate": ("INT", {"default": 2, "min": 1, "max": 100, "step": 1}),
                "black_point": ("FLOAT", {"default": 0.01, "min": 0.01, "max": 0.98, "step": 0.01}),
@@ -267,6 +267,18 @@ class ProcessDetails_RBS:
                local_files_only = detail_method == "VITMatte(local)"
                trimap = generate_VITMatte_trimap(orig_mask, detail_erode, detail_dilate)
                processed_mask = generate_VITMatte(orig_image, trimap, local_files_only, device, max_megapixels)
+            elif detail_method == "PyMatting":
+                trimap = generate_VITMatte_trimap(orig_mask, detail_erode, detail_dilate)
+                try:
+                    from pymatting import estimate_alpha_lkm
+                except ImportError:
+                    raise RuntimeError("请先安装 pymatting 库: pip install pymatting scikit-image")
+                import numpy as np
+                image_np = np.array(orig_image.convert('RGB'))
+                trimap_np = np.array(trimap.convert('L')) / 255.0
+                # PyMatting只用LKM算法，参数风格与LayerStyle_Advance一致
+                alpha = estimate_alpha_lkm(image_np, trimap_np)
+                processed_mask = Image.fromarray((alpha * 255).astype(np.uint8))
            else:  # GuidedFilter
                processed_mask = mask_edge_detail(i, m, detail_erode, black_point, white_point)
                processed_mask = tensor2pil(processed_mask)
@@ -1,9 +1,9 @@
 [project]
 name = "removebackgroundsuite"
 description = "A matting toolkit based on ComfyUI, supporting multiple matting models and detail processing methods."
-version = "1.0.0"
+version = "1.1.0"
 license = {file = "LICENSE"}
-dependencies = ["torch>=2.0.0", "numpy>=1.24.0", "Pillow>=9.0.0", "torchvision>=0.15.0", "opencv-python>=4.8.0", "scipy>=1.10.0", "transformers>=4.30.0", "huggingface-hub>=0.16.0", "tqdm>=4.65.0"]
+dependencies = ["torch", "numpy", "Pillow", "torchvision", "opencv-python", "scipy", "transformers", "tqdm"]

 [project.urls]
 Repository = "https://github.com/whmc76/ComfyUI-RemoveBackgroundSuite"
@@ -1,9 +1,8 @@
-torch>=2.0.0
-numpy>=1.24.0
-Pillow>=9.0.0
-torchvision>=0.15.0
-opencv-python>=4.8.0
-scipy>=1.10.0
-transformers>=4.30.0
-huggingface-hub>=0.16.0
-tqdm>=4.65.0 
+torch
+numpy
+Pillow
+opencv-python
+scipy
+torchvision
+transformers
+tqdm