Introduction
This project is an advanced web application that leverages the power of computer vision and deep learning to detect and classify human emotions in real-time. The system is designed with a clean, user-friendly interface that allows users to analyze emotions from three different sources: static images, pre-recorded videos, and a live webcam feed. By processing visual data, the trained machine learning model identifies faces and accurately predicts emotions such as happy, sad, fear, and neutral, overlaying the results directly onto the media for clear visualization.
Key Features
-
Multi-Modal Emotion Analysis: The application offers three distinct modes for emotion detection to cover various use cases:
-
Predict from Image: Upload a static image (e.g., JPG, PNG) to detect and label the emotion on each face within the picture.
-
Predict from Video: Upload a video file (e.g., MP4) to process it frame-by-frame, tracking and displaying the emotions of individuals as they change over time.
-
Real-Time Prediction: Activate the user's webcam for a live, real-time analysis of facial expressions, providing instant feedback.
-
-
Deep Learning Model: At its core is a trained neural network capable of recognizing key facial features and classifying them into distinct emotional categories.
-
Intuitive Web Interface: A simple and clean frontend allows users to easily select their desired mode, upload files, and start the prediction process with just a few clicks.
-
Visual Feedback with Bounding Boxes: The application draws bounding boxes around detected faces and labels them with the predicted emotion, offering clear and immediate results directly on the image or video feed.
Technology Stack
-
Backend & AI: Python
-
Web Framework: Flask / Django (to serve the model and UI)
-
Computer Vision: OpenCV
-
Deep Learning: TensorFlow / Keras / PyTorch
-
Frontend: HTML5, CSS3, JavaScript
Category
-
Artificial Intelligence / Machine Learning
-
Computer Vision
-
Web Application
