ComfyUI_FaceAlignPaste

智能人脸对齐与粘贴节点集合 | Smart Face Alignment and Pasting Node Collection

这是一个用于 ComfyUI 的基础人脸处理节点集合，提供智能人脸对齐、粘贴和自动适配功能，让您的人物图像处理工作流程更加高效和精准。

This is an advanced face processing node collection for ComfyUI, providing intelligent face alignment, pasting, and auto-fitting functions to make your character image workflow more efficient and precise.

---

功能特点 | Features

• 智能人脸检测 - 支持多种检测方法（OpenCV、MediaPipe、dlib），自动选择最佳结果
• 自动人脸对齐 - 根据人脸特征点智能对齐人脸，确保完美构图
• 精确定位系统 - 基于眼睛位置进行精确定位，实现专业级人像构图
• 智能缩放算法 - 根据人脸比例自动调整大小，保持人脸中心不变
• 多种构图模式 - 支持肖像、半身、全身等多种构图模式，满足不同创作需求
• 边界检查 - 自动检测并提示人物是否超出画布边界，避免意外裁剪
• 高质量图像处理 - 使用 Lanczos 插值等高质量算法确保图像清晰度

---

Smart Face Detection - Supports multiple detection methods (OpenCV, MediaPipe, dlib), automatically selecting the best result
Automatic Face Alignment - Intelligently aligns faces based on facial landmarks, ensuring perfect composition
Precise Positioning System - Positions based on eye location for professional portrait composition
Smart Scaling Algorithm - Automatically adjusts size based on face proportions while maintaining face center
Multiple Composition Modes - Supports portrait, half-body, full-body and other composition modes for different creative needs
Boundary Check - Automatically detects and alerts when characters exceed canvas boundaries to avoid accidental cropping
High-Quality Image Processing - Uses high-quality algorithms like Lanczos interpolation to ensure image clarity

---

节点说明 | Nodes Description

1. Face Align Single @ CHAOS

单人脸对齐节点，可以检测人脸并进行精确对齐处理，支持多种检测方法和对齐模式。

Single face alignment node that can detect faces and perform precise alignment, supporting multiple detection methods and alignment modes.

2. Face Align Double @ CHAOS

双人脸对齐节点，可以同时处理两张人脸图像并进行对齐，适用于人物对话场景或合影构图。

Double face alignment node that can process two face images simultaneously, suitable for character dialogue scenes or group photo compositions.

3. Face AutoFit Single @ CHAOS

智能人脸适配节点，可以将人脸图像智能地放置到指定尺寸的画布上，自动调整大小和位置，实现完美构图。主要参数包括构图模式、背景颜色、位置微调、人脸大小和旋转角度等。

Intelligent face auto-fitting node that can smartly place face images onto a canvas of specified size, automatically adjusting size and position for perfect composition. Main parameters include composition mode, background color, position fine-tuning, face size, and rotation angle.

4. Face AutoFit Double @ CHAOS

双人脸适配节点，可以同时处理两张人脸图像并放置在同一画布上，自动调整相对位置和大小，适合创建对话场景或人物互动图像。

Double face auto-fitting node that can process two face images simultaneously and place them on the same canvas, automatically adjusting relative position and size, ideal for creating dialogue scenes or character interaction images.

5. Image Rotate @ CHAOS

高质量图像旋转节点，提供精确的图像旋转功能，保持人脸中心不变，支持多种插值方法确保旋转后的图像质量。

High-quality image rotation node providing precise image rotation while maintaining face center, supporting multiple interpolation methods to ensure image quality after rotation.

---

安装方法 | Installation

1. 下载本仓库到 ComfyUI 的 custom_nodes 目录：

   Download this repository to the custom_nodes directory of ComfyUI:

cd "path_to_your_comfyui/ComfyUI/custom_nodes"
git clone https://github.com/CHAOSEA/ComfyUI_FaceAlignPaste.git

## 模型文件 | Model Files

本节点包含以下预训练模型文件：

- `models/shape_predictor_68_face_landmarks.dat` - dlib 人脸关键点检测模型，用于精确定位人脸的 68 个特征点

这些模型文件已包含在节点包中，无需单独下载。如果您遇到模型加载错误，请确保文件路径正确且文件完整。