IRJET - Emotion Recognising System-Crowd Behavior Analysis

This document describes an emotion recognition system that analyzes crowd behavior using machine learning and image processing techniques. The system works in three stages: 1) face detection using Haar cascade algorithms, 2) feature extraction and emotion recognition by converting images to grayscale and using CNN models, and 3) sending alerts based on recognized emotions like anger. The system was able to accurately detect faces and recognize emotions like happy, sad, anger, etc. It sent alerts via Twilio if high levels of anger were detected, allowing for analysis of crowd behavior and monitoring of public safety.