Core Research Specialization

Video Analytics

Temporal analysis of video data including action recognition, tracking, and video understanding.

2 Researchers 1 Projects 1 Publications

Collaborate in this Area Join Research Lab

In-Depth Overview

Research Focus & Scope

Video Analytics extends computer vision into the temporal dimension. ICVRL develops real-time video intelligence pipelines that understand actions, track multiple interacting targets, recognize behaviors, and flag anomalous events in live camera streams.

Methodological Approach

How We Conduct Research in Video Analytics

We utilize 3D convolutions, spatio-temporal graph neural networks, optical flow transformers, and multi-object tracking (MOT) frameworks with Kalman filtering and deep appearance embeddings for real-time video stream processing.

Impact & Deployments

Real-World Applications

Public safety drowning prevention and distress monitoring

Crowd density management and stampede prevention

Sports analytics and athlete performance tracking

Traffic flow optimization and congestion detection

Workplace safety protocol compliance monitoring

Lab Projects

Projects in this Domain

All Projects

ActionNet: Video Understanding for Smart Surveillance

Intelligent video analytics system for automated activity detection and anomaly recognition in surveillance fo...

active Code

Scholarly Output

Recent Publications

All Publications

conference

Temporal Action Detection in Surveillance Videos Using Graph Neural Networks

Hassan Ahmed, Muhammad Hamza, Shams ur Rehman

European Conference on Computer Vision (ECCV), 2024

Tech Stack & Tools

ByteTrack DeepSORT SlowFast VideoMAE OpenCV FFmpeg GStreamer PyTorch

Active Researchers

Shams Rehman

Director & Senior Lecturer

Hassan Ahmed

Undergraduate Researcher

View All Researchers

Join Our Research

Interested in conducting MS/PhD research or undergraduate projects in Video Analytics?

Apply to Join ICVRL

More Domains

Explore Other Research Areas

All Areas