SSTV (Smart Surveillance with Tracking and Vision)

SSTV is a comprehensive computer vision toolkit for video surveillance, human behavior analysis, and object tracking using YOLOv8 and machine learning technologies.

⚠️ Project Status: Archived
This project is no longer in active development. It has been open-sourced for reference and educational purposes. Community contributions and forks are welcome.

📝 Note: This README was automatically generated by Claude AI and may contain inaccurate information.

🎯 Features

• Real-time Object Detection: YOLOv8-based detection for persons, objects, and pose estimation
• Human Behavior Analysis: ML-powered behavior classification and tracking
• Video Labeling Tool: Interactive WPF application for manual video annotation
• Multi-tracking Support: ByteTrack integration for object tracking across frames
• Pose Estimation: Human pose detection and analysis
• Video Processing Pipeline: Automated video preprocessing and analysis
• Machine Learning Training: Custom model training capabilities

🏗️ Project Structure

The solution consists of several interconnected projects:

Core Libraries

• SSTV-Shared: Common models, utilities, and shared functionality
• SSTV-Shared-WPF: WPF-specific components and rendering utilities
• SSTV-MLTasks: Machine learning pipeline and YOLO processing engine

Applications

• SSTV-Client: Real-time surveillance application with Windows Forms UI
• SSTV-Labeler: Video annotation tool for creating training datasets
• SSTV-ML: Machine learning model training application
• SSTV-MLTasks-CLI: Command-line interface for batch processing
• SSTV-Benchmark: Performance benchmarking utilities

🚀 Quick Start

Prerequisites

• .NET 8.0 or later
• Windows OS (for WPF applications)
• Visual Studio 2022 or Visual Studio Code
• ONNX Runtime
• OpenCV

Installation

1. Clone the repository:

   git clone https://github.com/penspanic/SSTV.git
   cd SSTV

2. Open the solution in Visual Studio:

   cd src/SSCSharp
   start SSTV.sln

3. Build the solution:

   dotnet build

Running Applications

SSTV-Client (Real-time Surveillance)

cd src/SSCSharp/SSTV-Client
dotnet run

SSTV-Labeler (Video Annotation Tool)

cd src/SSCSharp/SSTV-Labeler
dotnet run

SSTV-MLTasks-CLI (Command Line Processing)

cd src/SSCSharp/SSTV-MLTasks-CLI
dotnet run -- video preprocess -i "./data/samples"

📚 Key Components

YOLOv8 Integration

The project uses YOLOv8 for object detection and pose estimation:

• Object detection models (yolov8s.onnx, yolov8m.onnx)
• Pose estimation models (yolov8s-pose.onnx)
• Custom model conversion scripts

Video Processing Pipeline

• Frame extraction and preprocessing
• Real-time processing capabilities
• Batch processing support
• Timeline-based event management

Machine Learning Features

• Human behavior classification
• Custom model training with ML.NET
• Training data management and serialization
• Performance benchmarking

🔧 Configuration

Model Setup

Place your ONNX models in the src/SSCSharp/Models/ directory:

• yolov8s.onnx - Standard object detection
• yolov8m.onnx - Medium accuracy object detection
• yolov8s-pose.onnx - Pose estimation

Data Directory Structure

Create the following directory structure for your data:

SSTV/
├── data/
│   ├── samples/          # Place your video files here
│   ├── input/            # Input datasets for ML training
│   └── output/           # Generated models and results
└── src/                  # Source code

FFMPEG Installation (Required for SSTV-Labeler)

SSTV-Labeler requires FFMPEG for video processing:

1. Download FFMPEG from https://ffmpeg.org/download.html
2. Extract and add to your system PATH, or place the binaries in the SSTV-Labeler/FFMPEG/ directory
3. Restart your development environment

📖 Documentation

For detailed documentation on each component:

• SSTV-Labeler README

🤝 Contributing

Since this is an archived project, we encourage forks and community-driven development:

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

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🏷️ Technologies Used

• .NET 8.0: Core framework
• YOLOv8: Object detection and pose estimation
• OpenCV: Computer vision operations
• ML.NET: Machine learning model training
• WPF: Windows desktop applications
• Windows Forms: Real-time surveillance interface
• ONNX Runtime: Model inference
• ByteTrack: Multi-object tracking
• FFMPEG: Video processing (via FFME.Windows)

⚠️ Known Issues

• Backward video navigation by 1 frame may not work properly due to FFMPEG library limitations
• Performance may vary depending on hardware capabilities
• Hardcoded paths in some configuration files need to be updated for your environment

🔮 Future Enhancements (Community Welcome)

• Cross-platform support
• Web-based annotation interface
• Enhanced model training pipelines
• Real-time streaming capabilities
• Advanced analytics dashboard
• Docker containerization

📞 Support

If you encounter any issues or have questions, please:

1. Check the Issues page
2. Create a new issue with detailed information
3. Provide system information and error logs

Since this is an archived project, community support is encouraged through discussions and community-maintained forks.

---

Made with ❤️ for computer vision and surveillance applications