Skip to main content
EmoChat Hero

​
Welcome to EmoChat

EmoChat is an AI-powered emotion recognition system that analyzes facial expressions in real-time to detect and understand human emotions. Built with OpenCV, scikit-learn, and Google Gemini AI, it provides empathetic feedback and emotional insights.

​
Key Features

Real-time Detection

Analyze emotions through webcam with instant facial landmark detection

Machine Learning

Random Forest classifier trained on facial expression patterns

Privacy-Focused

No data storage - all processing happens locally in real-time

Empathetic AI

Google Gemini AI provides supportive emotional analysis

​
What You Can Do

EmoChat enables you to:
  • Detect Emotions: Recognize Happy and Sad emotions from facial expressions
  • Track Sessions: Record 30-second emotion tracking sessions
  • Get Insights: Receive empathetic feedback from AI based on emotional patterns
  • Train Custom Models: Prepare and train your own emotion recognition models
  • Build Applications: Integrate emotion detection into web applications

​
How It Works

1

Facial Landmark Detection

OpenCV detects faces and extracts 68 facial landmark points using Haar Cascade and LBF models
2

Feature Normalization

Landmark coordinates are normalized relative to face boundaries for consistency
3

ML Classification

Random Forest classifier analyzes normalized features to predict emotions
4

Real-time Feedback

Results are displayed instantly with optional AI-powered emotional analysis

​
Use Cases

Track emotional patterns during therapy sessions or personal reflection exercises
Analyze user reactions to products, interfaces, or content in real-time
Learn about computer vision, machine learning, and emotion recognition systems
Build applications that respond to users’ emotional states

​
Quick Start

Get up and running in minutes:

Quickstart Guide

Launch the web app and start detecting emotions

Installation

Install dependencies and set up your environment

Train Your Model

Prepare data and train custom emotion recognition models

API Reference

Explore the core functions and HTTP endpoints

​
Technology Stack

EmoChat is built with:
  • OpenCV: Face detection and landmark extraction
  • scikit-learn: Random Forest machine learning classifier
  • Flask: Web server and REST API
  • Google Gemini AI: Empathetic emotional analysis
  • JavaScript: Real-time webcam integration
All emotion processing happens locally on your machine. No facial data is stored or transmitted to external servers (except when using the optional Gemini AI analysis feature).

​
Community & Support

  • GitHub Repository: Dani-zm/emotion-recognition-ai
  • Issues: Report bugs or request features on GitHub
  • Documentation: Comprehensive guides and API reference
Ready to get started? Head to the Quickstart Guide to launch your first emotion detection session.

Build docs developers (and LLMs) love

Get started for freeTalk to us