ComfyUI-OpenPose-Keypoint-Extractor

★ 36

人体姿态边界框生成OpenPose关键点处理

解析OpenPose的POSE_KEYPOINT输出，结合原图宽高计算各骨骼部位的(x,y,width,height)边界框，支持填充以覆盖完整肢体，便于裁剪、检测与后处理。

💡 将OpenPose关键点转换为肢体/面部的裁剪或检测边界框

🍴 10 Forks💻 Python🔄 2025-07-30

🔗 GitHub 原文

📦

网盘下载

复制链接后前往夸克网盘下载

https://pan.quark.cn/s/9671236b7e59

📄 README

Openpose Keypoint Extractor

Take the keypoint output from OpenPose estimator node and calculate bounding boxes around those keypoints.

You need to give it the width and height of the original image and it will output (x,y,width,height) bounding box within that image

Note that the points on the OpenPose skeleton are *inside* the particular limb (eg center of wrist, middle of shoulder), so you probably will want to apply some padding around the bounding box to get the whole arm, leg or whatever you’re looking for.

Reference for which control point is which – provide comma-separated list in the Node: OpenPose Docs

Example workflow

Drag this to your ComfyUI: