Publication Details

Abstract

Crowded public environments such as hospitals, transportation hubs, and service centers are increasingly affected by long waiting times and overcrowding, posing significant safety and operational challenges. This thesis presents CrowdIQ, an AI-powered system for real-time monitoring and intelligent management of crowded areas. The proposed solution integrates computer vision, and machine learning, to process live video feeds through a multi-stage pipeline involving preprocessing, object detection, and queue analysis. OLO and convolutional neural networks (CNN) are used for object detection and emotion classification. The system focuses on real-time monitoring, queue estimation, and crowd analysis, with potential extensions including predictive analytics and anomaly detection Experimental evaluation demonstrates improved accuracy in crowd analysis and enhanced operational efficiency. The proposed system is scalable across multiple domains and contributes to improved safety, reduced waiting times, and better service delivery.