A Robust Deep Networks based Multi-Object MultiCamera Tracking System for City Scale Traffic
Muhammad Imran Zaman, Usama Ijaz Bajwa, Gulshan Saleem, Rana Hammad Raza
2025-05-02
Summary
This paper talks about a powerful AI system that can keep track of lots of different vehicles at once using footage from multiple cameras all over a city.
What's the problem?
It's really hard to accurately follow many moving vehicles in busy city traffic, especially when using video from different cameras that might have different angles, lighting, or quality.
What's the solution?
The researchers combined several advanced AI techniques, like deep learning models and smart tracking algorithms, to create a system that can recognize and follow vehicles across different cameras, even in tough conditions.
Why it matters?
This matters because it can help make city traffic safer and more efficient by allowing authorities to monitor traffic flow, spot accidents, and manage congestion more effectively.
Abstract
A deep learning framework using Mask R-CNN, transfer learning, ResNet-152, and Deep SORT achieves competitive performance in multi-object, multi-camera vehicle tracking under diverse urban traffic conditions.