YOLOv8 Pose and Object LSTM Bluetooth Combined

Introduction

This project integrates YOLOv8 for pose detection and object recognition, LSTM (Long Short-Term Memory) for sequential analysis, and Bluetooth communication for real-time interaction. It is designed to provide a robust solution for tracking and analyzing pose and object data in dynamic environments.

Quick Start

This is my HackMD.io post about the YOLOv8 Tutorial. It serves as a quick guide to help you get started efficiently. The content is written in Chinese and is user-friendly for Chinese-speaking readers. Feel free to refer to it for your learning needs!

---

Features

• Pose Detection: Utilizing YOLOv8 for real-time pose estimation.
• Object Recognition: Detects objects in frames with high accuracy.
• Sequential Analysis: Implements LSTM to analyze pose and object sequences for pattern recognition.
• Bluetooth Integration: Enables wireless data transfer for real-time applications.
• Customizable and Scalable: Easily adaptable for different use cases, including sports analytics, robotics, and more.

---

TODO List: Bluetooth Integration

This section outlines the tasks required to integrate Bluetooth communication into the project.

1. Bluetooth Module Configuration

   • Install and configure the Bluetooth module on your device (e.g., HC-05, HC-06, or similar).
   • Could you verify the module is discoverable and properly paired with your computer or target device?

2. Connection Protocol

   • Define the communication protocol for transferring pose and object data (e.g., JSON or CSV format).
   • Implement handshaking to ensure reliable data transfer between the device and the application.

3. Bluetooth API

   • Utilize PyBluez or an equivalent library for Python-based Bluetooth communication.
   • Implement basic Bluetooth functions:
     • Device scanning
     • Pairing and unpairing
     • Sending and receiving data

4. Integration with YOLOv8 and LSTM

   • Update the main.py script to send YOLOv8 and LSTM outputs to the paired Bluetooth device.
   • Design a feedback mechanism to receive acknowledgment or additional commands via Bluetooth.

5. Testing and Debugging

   • Test Bluetooth communication in a controlled environment with static and dynamic data.
   • Debug any latency or data loss issues during transfer.

6. Documentation

   • Document all configurations, including device names, UUIDs, and pairing codes.
   • Provide a user guide for setting up the Bluetooth environment.

---

Installation

To set up the project, follow these steps:

1. Clone the repository:

   git clone https://github.com/Yucheng0208/YOLOv8_Pose_and_Object_LSTM_Bluetooth_Combined.git
   cd YOLOv8_Pose_and_Object_LSTM_Bluetooth_Combined

2. Create a Python virtual environment (optional but recommended):

   python -m venv venv
   source venv/bin/activate  # For Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Set up the environment:

   • Ensure that your device installs and configures the necessary Bluetooth modules.
   • Update the configuration file config.yaml with your device-specific parameters.

---

Usage

1. Run the YOLOv8 pose and object detection module:

   python yolo_detection.py

2. Enable LSTM sequential analysis:

   python lstm_analysis.py

3. Start Bluetooth communication for data transfer:

   python bluetooth_transfer.py

4. Combine all modules:

   python main.py

---

File Structure

YOLOv8-Enhanced-LSTM-BT/
│
├── data/                   # Sample datasets and testing data
├── models/                 # YOLOv8 and LSTM pre-trained models
├── scripts/                # Custom utility scripts
├── yolo_detection.py       # YOLOv8 pose and object detection
├── lstm_analysis.py        # LSTM sequential analysis
├── bluetooth_transfer.py   # Bluetooth communication script
├── main.py                 # Main program combining all modules
├── config.yaml             # Configuration file for parameters
└── requirements.txt        # Required Python libraries

Citation

@inproceedings{tsai2024based,
  title={Using Real-Time Integrated Computer Vision and Deep Learning for Advanced Factory Safety},
  author={Tsai, Shang-Te and Chang, Yu-Cheng and Huang, Chao-Yi and Yang, Fang-Yuan and Chien, Wei and Wu, Zong-Rong},
  booktitle={2024 IEEE 4rd International Conference on Electronic Communications, Internet of Things and Big Data (ICEIB)},
  pages={647-649},
  year={2024},
  organization={IEEE},
  doi={10.1109/ICEIB61477.2024.10602722}
}}

---

License

This project is licensed under the MIT License.