A full-stack application built with EmguCV, C# .NET MVC, and JavaScript, enabling real-time pose detection on images and videos. Users can visualize human keypoints, manually correct poses, and download updated media for fitness, sports analysis, or experimentation.
This repository is actively maintained and will be continuously updated with new features, improvements, and experiments.
- Image Pose Detection: Detect keypoints on uploaded images and receive annotated outputs.
- Video Pose Detection: Process video files frame-by-frame with overlayed pose results.
- Interactive Pose Correction: Drag and adjust keypoints using a canvas-based UI.
- Export Functionality: Download corrected images or videos with new pose layouts.
- Modular Codebase: Easily extendable and adaptable for other computer vision projects.
- Frontend: HTML5, JavaScript (Canvas API)
- Backend: C# (.NET MVC), EmguCV (OpenCV wrapper)
- Libraries: EmguCV, jQuery
- .NET Core 7
- EmguCV installed
- Visual Studio 2022 or later
- Clone the repository:
git clone https://github.com/arooshahmad-data/pose-estimation-correction-ui-emgucv.git
- Open the .sln file in Visual Studio.
- Restore NuGet packages.
- Build and run the application.
- Upload an image or video.
- The system detects pose keypoints using EmguCV.
- Interactively correct poses via the drag-based canvas UI.
- Download the updated media with corrected poses.
/Controllers # MVC Controllers
/Views # Razor Views
/Scripts # JavaScript & canvas logic
/Models # C# Models for data exchange
/docs/result_images # Output preview samples
/wwwroot # Static files (JS, CSS, assets)
This repository is continuously updated. Contributions, suggestions, and improvements are welcome to enhance features or add new pose detection experiments.
Author: Aroosh Ahmad — AI Engineer (NLP, LLMs, ML Systems) GitHub • LinkedIn