BCA Thesis Project

Real-Time Emotion Detector

Facial emotion recognition using deep learning. Detects happiness, sadness, anger, surprise, and more in real time through your webcam.

Loading AI models...

About This Project

Architecture

Uses TinyFaceDetector (a lightweight CNN) for real-time face detection and a deep neural network trained on the FER-2013 dataset for expression classification into 7 emotion categories.

Hardware Used (Original)

Raspberry Pi 3B+, USB webcam, Bluetooth microphone, HDMI display. The original BCA thesis demo ran on embedded hardware with voice command activation using Google Text-to-Speech.

Tech Stack

Python, OpenCV (Haar Cascades), TensorFlow (MobileNet), Tkinter GUI, gTTS for voice commands. This web demo uses face-api.js with TensorFlow.js for browser-native inference.

Privacy Note

All processing happens locally in your browser. No video or image data is sent to any server. The camera feed never leaves your device.