NeuroAI Backend

This project is part of the NeuroAI platform, a multi-modal AI-powered backend designed to support psychiatric diagnosis through emotion recognition and subconscious analysis.
This project was developed as part of the coursework for ESE.INFIA0010 - AI Project at Esprit School of Engineering.

Overview

NeuroAI Backend implements a multi-container microservices architecture in Docker to process real-time emotion detection and brain-to-text insights using Deep Learning models.
Each microservice is responsible for a distinct modality (EEG&EOG, EEG during activity, ECG, speech, facial, Brain-to-text), and all data is aggregated via a gateway for streamlined communication.

Features

• Multi-modal approach to emotion recognition
• Dockerised Microservice architecture for portability, scalability and flexibility
• Supports various emotion recognition modalities:
  • 🎤 Speech Emotion Recognition
  • 👤 Facial Emotion Recognition
  • 🧠 EEG&EOG Emotion Interpretation
  • ❤️ ECG-Based Emotional Analysis
  • 🕹️ Gaming-Stimulated EEG Emotion Tracking
• 🧠 Brain-to-text interpretation using EEG signals

Tech Stack

• Backend: Python, Flask.
• AI/ML: TensorFlow, PyTorch, Deep Learning.
• Infrastructure: Docker, Docker Compose, Reverse Proxy Gateway.
• Other Tools: Postman, Jupyter Notebooks, ngrok.

Directory Structure

root/
│
├── microservices/
│ ├── microservice_a/ # One use case (e.g. speech)
│ ├── microservice_b/ # Another use case (e.g. facial)
│ └── ...
│
├── gateway/ # API gateway (user entry point)
├── docker-compose.yml # Runs all microservices together
└── README.md

Getting Started

1. Clone the repository.
2. Build and run all services: docker-compose up --build
3. Use the API Gateway to make requests.

Usage

The gateway is a single entry point. It receives user requests and redirects them to the right microservice based on the endpoint path.

Contributing

Check out CONTRIBUTING.md for guidelines.

Acknowledgments

This project was developed as part of the coursework for AI project (ESE.INFIA0010) at Esprit School of Engineering.

Special thanks to

• Prof. Sonia Mesbeh (sonia.mesbeh@esprit.tn)
• Prof. Jihene Hlel (jihene.hlel@esprit.tn)

for their invaluable guidance from ideation and model training through to system architecture and deployment.

Topics

• artificial-intelligence
• deep-learning
• machine-learning
• data-analysis
• web-development
• python
• flask
• API
• docker
• containerization