Robust, Extensible, and Fast: Teamed Classifiers for Vehicle Tracking and Vehicle Re-ID in Multi-Camera Networks

TL;DR

This paper introduces a robust, extensible, and real-time teamed classifier framework for vehicle tracking and re-identification in heterogeneous multi-camera networks, addressing challenges like variability, occlusion, and adversarial conditions.

Contribution

It presents a novel teamed classifier system that operates in real-time, handling diverse camera setups and environmental challenges for vehicle analytics.

Findings

Robust performance on VeRi-776 and Cars196 datasets.

Extensible to new vehicle types and camera configurations.

Achieves real-time processing compared to offline methods.

Abstract

As camera networks have become more ubiquitous over the past decade, the research interest in video management has shifted to analytics on multi-camera networks. This includes performing tasks such as object detection, attribute identification, and vehicle/person tracking across different cameras without overlap. Current frameworks for management are designed for multi-camera networks in a closed dataset environment where there is limited variability in cameras and characteristics of the surveillance environment are well known. Furthermore, current frameworks are designed for offline analytics with guidance from human operators for forensic applications. This paper presents a teamed classifier framework for video analytics in heterogeneous many-camera networks with adversarial conditions such as multi-scale, multi-resolution cameras capturing the environment with varying occlusion, blur, and orientations. We describe an implementation for vehicle tracking and vehicle re-identification (re-id), where we implement a zero-shot learning (ZSL) system that performs automated tracking of all vehicles all the time. Our evaluations on VeRi-776 and Cars196 show the teamed classifier framework is robust to adversarial conditions, extensible to changing video characteristics such as new vehicle types/brands and new cameras, and offers real-time performance compared to current offline video analytics approaches.