Pose-Evaluation System

An intelligent desktop application for pose evaluation and visualization. The UI is built on the refined PyQt5 component library QFluentWidgets, delivering a modern Fluent design experience. Local LLM assistance (via Ollama) transforms numeric scores into actionable guidance.

🌟 Overview

Pose-Evaluation System provides an end-to-end workflow for ingesting videos, analyzing movements, visualizing per-limb metrics, and generating tailored improvement suggestions with the help of a local LLM. It is engineered with PyQt5 and QFluentWidgets to achieve a clean, responsive, and consistent desktop UX. This is a project extended from College Students' innovation and Entrepreneurship Competition of China, with the help of H. Huang, under the supervision of H. Wang.

✨ Key Features

📥 Video Ingestion

• Load both standard reference videos and user videos for analysis
• Built-in demo assets for quick exploration

📊 Pose Evaluation & Visualization (Vis)

• Overall and per-limb scoring (e.g., standardness/accuracy and amplitude)
• Visualization pages to inspect results at a glance

🎞️ Video Results (VID)

• Play aligned/processed comparison videos for qualitative inspection

🤖 AI Assistant

• Local LLM (Ollama) turns computed metrics into human-friendly recommendations
• Default model: qwen2.5:3b (configurable)

⚙️ Settings & 🔎 Help

• Theme, language, and download directory management
• Help and support links surfaced within the app

🚀 Quick Start

Prerequisites

• Python 3.11+
• Windows 10/11 recommended (cross-platform PyQt is supported)

Installation

git clone <your-repo-url>
cd Pose-Estimation System
pip install -r requirements.txt

(Optional) Local LLM with Ollama

# Install and start Ollama (see official docs)
ollama pull qwen2.5:3b
# If running remotely, set OLLAMA_HOST accordingly

Launch

python main.py

On first launch, the application reads app/common/config.py and app/config/config.json for theme, language, and folder settings.

📖 Usage Guide

1. Home: read quick guidance; Help: access support links
2. Demo/Upload: import a standard reference video and a user video
3. Vis: review overall and per-limb scores with visualizations
4. VID: play the aligned result video
5. AI: ask domain questions and receive tailored improvement suggestions

🏗️ Architecture

Frontend (Desktop UI)

• PyQt5 + PyQtWebEngine: core UI and embedded web content
• QFluentWidgets: modern Fluent design components and UX patterns

Views & Navigation

• MainWindow organizes pages: Home, Upload, Vis, VID, AI, Help, Settings
• Reusable components located under app/components/*

Common Utilities & Resources

• app/common/* for config, styling, signals, i18n, resource loading
• app/resource/* for images, icons, QSS themes, and translations managed by resource.qrc

Optional Local LLM

• app/view/llm_worker.py integrates Ollama for on-device inference
• qwen2.5-7b/* shows a Transformers/LangChain-based local LLM class for experimentation

🔧 Configuration

• Runtime configuration is stored in app/config/config.json and described in app/common/config.py
• Key options: theme mode, DPI scaling, language, download folder
• Update URLs for Help/Home in app/common/config.py as needed

📁 Project Structure

root
│  main.py                 # Application entry point
│  run_server.py           # Placeholder for server-side extensions
│  requirements.txt        # Python dependencies (replaces PESUI_ENV.yml)
│
├─app
│  ├─common                # Config, style, signals, i18n, resources
│  ├─components            # Reusable UI components
│  ├─resource              # Images / QSS themes / translations / qrc
│  ├─UI                    # QtDesigner generated and adjusted UI
│  └─view                  # Page views and navigation
│
└─qwen2.5-7b              # Optional: local LLM examples (Qwen)
       llm.py
       api-chat.py
       download_model.py

🤝 Contributing

1. Fork the repository
2. Create a feature branch: git checkout -b feature/AmazingFeature
3. Commit your changes: git commit -m "Add AmazingFeature"
4. Push to the branch: git push origin feature/AmazingFeature
5. Open a Pull Request

📄 License

Specify the license that applies to this project and add a LICENSE file. Third-party dependencies are subject to their respective licenses.