VisualPosture
AI Fitness Coach with BlazePose: Real-Time Exercise Feedback This project implements an AI-powered fitness coaching system using MediaPipe's BlazePose and OpenCV for real-time human pose estimation. Designed to address the lack of automated exercise guidance
Install / Use
/learn @22xxlin/VisualPostureREADME
AI健身动作分析系统 / AI Fitness Action Analysis System

视频演示 / Video Demo
点击上方GIF观看完整演示视频
Click the GIF to watch full demo video
项目概述 / Overview
中文
基于PyQt5和MediaPipe BlazePose实现的实时健身动作分析系统,具备:
- 33个人体关键点实时追踪
- 右膝关节动作完成度可视化
- 视频/摄像头多源输入支持
- 多线程同步控制架构
English
Real-time fitness action analysis system based on PyQt5 and MediaPipe BlazePose, featuring:
- Real-time tracking of 33 body keypoints
- Visualized completion rate for right knee joint
- Multi-source input (video/camera)
- Multithreading synchronization architecture
核心功能 / Core Features
动作分析 / Action Analysis
| 中文描述 | English Description | |---------|---------------------| | 24-26-28关键点角度计算 | Angle calculation for keypoints 24(hip)-26(knee)-28(ankle) | | 120°→170°弯曲映射为0-100%进度条 | 120°→170° bending mapped to 0-100% progress bar | | 动态颜色反馈(洋红→绿) | Dynamic color feedback (magenta→green) |
交互功能 / UI Controls
| 按钮 | Button | 功能 | Function | |-----|--------|-----|----------| | 打开文件 | Open File | 选择视频文件 | Select video file | | 暂停/继续 | Pause/Resume | 控制分析流程 | Control analysis process | | 退出 | Exit | 关闭应用 | Close application |
环境配置 / Environment Setup
依赖要求 / Requirements
# 中文版本
Python == 3.7+ # 解释器
opencv-python == 4.5.2 # 图像处理
mediapipe == 0.8.3 # 姿态检测
PyQt5 == 5.15.4 # 图形界面
numpy == 1.21.6 # 数值计算
# English Version
Python == 3.7+ # Interpreter
opencv-python == 4.5.2 # Image processing
mediapipe == 0.8.3 # Pose detection
PyQt5 == 5.15.4 # GUI framework
numpy == 1.21.6 # Numerical computing
使用指南 / Quick Start
启动命令 / Launch
python youcandoit.py
操作流程 / Workflow
1点击"打开视频"选择训练录像
1Click "Open File" to select workout video
2观察右侧膝关节进度条变化
2Monitor right knee progress bar (right-side vertical bar)
3使用暂停按钮进行动作分析
3Use pause button for detailed motion analysis
📜 References
- Blog Post: 基于BlazePose的AI健身教练系统
- Video Example:BlazePose+Pyqt5+OpenCV=>基于人体姿态识别的AI健身系统
- MediaPipe Documentation: BlazePose Overview
- PyQt5 Framework: Official Documentation

