Facial Profiling Pipeline

This project implements a real-time facial recognition system integrated with servo-controlled camera tracking (Feetech STS3032).

Features

• Real-time face detection and tracking
• Servo-controlled camera movement (currently commented out)
• Feature vector extraction for facial recognition
• Pinecone integration for vector similarity search
• MongoDB integration for user profiles
• Asynchronous processing for improved performance

Quick Start

1. Clone the repository and install dependencies:

   git clone https://github.com/krishgoel/realtime-facial-profiling.git
   cd facial-profiling-pipeline
   pip install -r requirements.txt
   

2. Set up environment variables in a .env file:

   PINECONE_API_KEY=your_pinecone_api_key
   MONGO_URI=your_mongodb_uri
   

3. Run the application:

   python -m app
   

Servo Control and Mapping

This project uses Feetech STS3032 servos for camera pan and tilt control. Two servos are used -

1. Pan servo: Controls horizontal movement
2. Tilt servo: Controls vertical movement

Note: The servo-controlled camera movement functionality is currently commented out in the code. Users may implement this feature as desired.

Top-View of Azimuthal Angle to Pan Servo Position Mapping (Servo ID 2). Axis DH aligns with the +ve X-axis, where vector AH faces the speaker. The green arc indicates the permissible panning range of the servo (-47.5° to 97.5°).

Side-View of the Elevation Angle to Tilt Servo Position Mapping (Servo ID 1). Axis AB aligns with the +ve X-axis where vector AB points to the speaker. The green arc indicates the permissible tilting range of the servo (-10° to +50°).

The servo mapping can be customized in the configuration file (config.py) to adjust the range and sensitivity of camera movement. This allows for fine-tuning the system to specific hardware setups and use cases.