OpenPose: 2D Pose Estimation

🚀 Real-Time Multi-Person 2D Pose Estimation Using OpenPose

This project implements OpenPose using deep learning for real-time multi-person 2D pose estimation with Part Affinity Fields (PAFs). It detects and tracks body key points from images and videos.

🚀 Live Notebook (Google Colab)

Run the project live without any setup:
🔗 Colab Notebook

📌 Project Overview

• Developed a multi-person pose estimation model using OpenPose and CMU’s Caffe-based framework.
• Implemented and tested in Google Colab for cloud-based execution.
• Applied Part Affinity Fields (PAFs) for pose tracking across frames.
• Evaluated results using custom datasets and pre-trained models.

📂 Project Structure

• code/ → Python scripts used for running OpenPose in Colab.
• datasets/ → Sample dataset used for training/testing.
• results/ → Output images/videos from the model.

🔗 Resources & References

• Official OpenPose Repository: CMU-Perceptual-Computing-Lab
• Research Paper: OpenPose: Realtime Multi-Person 2D Pose Estimation

📝 License

This project follows the MIT License.