Trash Throwing Action Detection

resultdemo.mp4

Overview

This project is designed to detect trash-throwing actions using computer vision techniques. It utilizes YOLOv8 for pose estimation to identify specific human actions related to throwing trash.

Features

• Pose Estimation: Uses YOLOv8 for detecting human poses.
• Action Detection: Identifies the action of throwing trash.
• Result Visualization: Provides visual outputs of the detection process.

Installation

1. Clone the repository:

   git clone https://github.com/bhaskarshukla002/Trash-Throwing-Action-Detection.git
   

2. Navigate to the project directory:

   cd Trash-Throwing-Action-Detection
   

3. Install the required dependencies:

   pip install opencv-python numpy torch ultralytics
   

Usage

1. Jupyter Notebook:

   • Open main.ipynb to run the detection process step by step.

2. Python Script:

   • Run the detection script: python python_implementation.py

Project Structure

• main.ipynb: Jupyter Notebook for interactive development and testing.
• python_implementation.py: Python script for action detection.
• result images/: Directory containing result images.
• videos/: Directory containing input video files.
• resultdemo.mp4: Demo video showcasing the detection results.
• yolov8n-pose.pt: Pre-trained model weights for YOLOv8 pose estimation.

Contributing

Contributions are welcome! Please submit a pull request or open an issue to discuss any changes.

License

This project is licensed under the MIT License.

Contact

For any inquiries, please contact the repository owner at [email protected].

---

For more details, visit the repository.