Mocap for All

Demo	Description
	Democratized motion capture via: Real Time Pose Transfer onto 3D Characters. An affordable motion capture system that uses real-time pose estimation to replace the expensive equipment. A comprehensive evaluation of state-of-the-art pose estimation algorithms and demonstrate their practical use in real-time 3D avatar animation required for traditional motion capture.

🔑 Key Feature

• A comprehensive comparison of multiple realtime pose estimation method
• A blender plugin to transfer your pose in real time to a 3D character
• Real time pose recording without motion capture suit

🧑‍💻️ Usage

Go into amazon_webscraper.py and do these changes

Open pose - pose estimation preview

Mask RCNN - pose estimation preview

Move Net - pose estimation preview

Media pipe - pose estimation preview

Pose Net - pose estimation preview

Mask RCNN - pose estimation preview

Step 6: PROFIT!!!

💻 Sample Output

🛠 Skills

Python , Computer Vision , Deep Learning , Blender , Unreal Engine , 3D model Rigging

📖 Libraries Used

Posenet, yolo , maskrcnn , openpose, media pipe, blaze pose

📃 Lessons Learnt

• Implementing multiple pose estimation libraries
• Difficulty of working with 3D rendering software

🔮 Future Scope

• Unreal Engine rigging to automate videogame motion capture
• Customized emojis using media-pipe face and hand tracking