Detection and monitoring of crowds using computer vision has applications in crowd management and surveillance. Crowd management is important for public safety, especially now amidst the COVID-19 pandemic. Computer vision algorithms can assist with social distancing efforts aimed at slowing the spread of the virus, and alert when violations on the permitted headcount within a space occur. Crowd counting and localization can also be useful when designing public spaces such as airports and malls, and in making decisions on how to manage crowds in these public spaces. Team Members Tasneem Naheyan Kenan Li Inder Dhillon Jing Li Sadman Sakib Reza Karimi Youssef Guirguis Data We used the WILDTRACK dataset for this project. Repo Structure Different components of the project are divided into seperate branches: wildtrack_dataset contains the homography transformation and distance calulation functions. YOLO branch contains the YOLOv3 model code. evaluation branch contains the code to run the evaluation metrics. Pre-generated YOLO detections provided in .pkl files. Kernel_Density_Estimation contains the KDE code. main branch contains the script to generate location position predictions and saves them to .pkl files.

Live human detection using Computer vision with OpenCV and TensorFlow to accurately detect and count people in real-time video streams, images, or live camera feeds. Analyze crowd density and generate detailed reports Visualize the data through dynamic graphs and export comprehensive PDF reports for further analysis and decision-making.

This project provides a real-time people counting system using computer vision. It tracks individuals in video streams, featuring centroid tracking, vertical line crossing detection, and confidence metrics. Built with Python and OpenCV, it’s scalable for crowd monitoring, security, and occupancy control.

Counting Faces Detector is a project designed to accurately count the number of faces in images and video feeds. Utilizing OpenCV for computer vision tasks and Python for scripting, this tool provides real-time face detection and counting capabilities. Ideal for security, surveillance, and crowd analysis applications.

Real-Time Human Counting and Detection Developed a real-time computer vision system to detect and count humans in video streams using Python, OpenCV, and deep learning. Features include bounding boxes, live counting, and data logging, useful for crowd monitoring, surveillance, and smart analytics.

"Number of People in a Room" is a computer vision project that utilizes the Flask framework, OpenCV library, and YOLO object detection to count and monitor people in a space. It aids in crowd management, occupancy monitoring, and optimizing resource allocation for improved safety and efficiency.

A smart crowd detection system that uses deep learning and computer vision to detect people in an image/video, estimate area density, and predict the risk of a stampede or need for crowd control. Combines YOLOv5 for object detection and custom logic for crowd risk analysis.

Real-time object detection and tracking using Computer Vision for detection and SORT for tracking. This repo processes video feeds to label moving objects with unique IDs. Ideal for traffic and crowd monitoring.

A Python project leveraging OpenCV and deep learning for real-time crowd detection and density estimation. Built for research and surveillance applications, it demonstrates practical implementation of computer vision models in safety and event-monitoring scenarios.

A real-time AI-powered person detection, tracking, and counting system using SSD MobileNet V3 and OpenCV. This project leverages deep learning and computer vision to automate crowd analytics for applications in smart cities, security, retail, and event management.

This project aims to enhance public safety, security and workplace efficiency by AI and ML with existing CCTV infrastructure. By leveraging computer vision and real-time analytics, the system will provide automated surveilance, anomaly detection, and predictive insights for crowd management, crime prevention, and workforce monitoring.

Crowd surveillance system developed for Prayagraj Mahakumbh Hackathon 2025. Utilizes advanced computer vision for head counting, anomaly detection, and fire detection in real-time, ensuring event safety and security. Proud finalists, dedicated to leveraging technology for positive impact.

A real-time surveillance system that detects suspicious activities and theft in crowded public spaces using YOLOv8 object detection. Integrated with Twilio for WhatsApp alerts and Google Gmail API for cloud notifications, this project offers instant visual proof and alerts to authorities — ensuring faster incident response and public safety.

Detection and monitoring of crowds using computer vision.

Computer Vision-based crowd event detection prototype for researches

AI-based crowd monitoring system for real-time density estimation and overcrowding detection using computer vision.

Comparative analysis of classical computer vision methods (HOG, SIFT, and SVM) for real-time crowd detection and crowd level classification using CCTV footage.

A web-based crowd detection system that uses computer vision and deep learning to detect and analyze crowds in videos. The system leverages YOLOv8 for person detection and implements custom logic for crowd analysis.

A web-based crowd detection system that uses computer vision and deep learning to detect and analyze crowds in videos. The system leverages YOLOv8 for person detection and implements custom logic for crowd analysis.

Real-time crowd risk detection system using Computer Vision to analyze density and motion anomalies for public safety.

Ezycrowd is an AI-driven smart crowd detection and management system leveraging computer vision to analyze real-time crowd density and generate alerts for improved public safety.

Real-Time Crowd Monitoring project utilizes AI and computer vision to analyze live video feeds for crowd density estimation, movement tracking, and anomaly detection. By integrating YOLO for object detection and Deep SORT for tracking, the system provides real-time insights, enhancing safety and crowd management.

Computer vision–based people counting system using Haar Cascade classifiers for real-time human detection. Implemented video frame processing and object detection pipelines with OpenCV, applying classical machine learning–based vision techniques for crowd analysis.

A computer-vision pipeline for real-time crowd density analysis, motion-aware hotspot detection, and early stampede risk warning from surveillance or event video.

Real-time Computer Vision Dashboard for traffic and crowd analytics. High-performance object detection and tracking (vehicles/pedestrians) using YOLOv8, ByteTrack, and OpenCV, served via a premium Streamlit UI.

Crowd and Face Mask Detection uses computer vision models to detect people and identify mask usage in images or video streams. It supports real-time inference, making it useful for monitoring crowds and safety compliance. The repo includes scripts for training, validation, detection, and pre-trained weights.

A Python-based crowd counting tool that utilizes computer vision to automatically detect and count people in images. Features include image preprocessing, head detection, and automated count reporting. Designed for analyzing high-density crowds in static photos with high accuracy.

To simulate an IoT-based smart surveillance system that estimates crowd density from real-time recorded videos using computer vision and deep learning. Students will collect their own data, annotate it, and train an AI model for people detection and crowd counting.

This project uses computer vision techniques to count people on a basketball court in images and videos. It employs deep learning models for accurate crowd detection , with visual bounding boxes around detected individuals. Ideal for learning image analysis and object detection.

AI-Based Real-Time Crowd Burst Detection System uses computer vision and machine learning to monitor live CCTV feeds, analyze crowd density and movement patterns, detect abnormal surges, and generate instant alerts. It transforms traditional surveillance into a proactive early-warning system for public safety.