Real-time crowd density estimation using CSRNet (VGG-16 + dilated CNN) on ShanghaiTech Part A & B. Includes adaptive Gaussian kernel density map generation, from-scratch training on M1 CPU, MPS inference, and a live video demo with heatmap overlay and running count graph. Part A MAE: 87.85 | Part B MAE: 21.12

A real-time computer vision system for monitoring and analyzing crowd density using YOLO-based object detection and tracking.

Crowd Stampede Analysis is a comprehensive computer vision pipeline for real-time crowd density estimation and stampede risk detection. It integrates YOLOv8 detection, DeepSORT tracking, motion analytics, and regression-based counting (CSRNet/MCNN) to process videos and generate density heatmaps, risk alerts, and performance reports.