CORT: Class-Oriented Real-time Tracking for Embedded Systems

TL;DR

This paper introduces CORT, a class-oriented real-time multi-object tracking method that improves timing predictability and efficiency in embedded systems without sacrificing tracking accuracy, especially in urban scenarios.

Contribution

The paper presents a novel multi-class tracking approach that reduces matching complexity by class-based splitting, enhancing real-time performance while maintaining accuracy.

Findings

Effective in urban multi-object scenarios

Outperforms state-of-the-art trackers in speed and accuracy

Maintains high tracking quality with predictable execution times

Abstract

The ever-increasing use of artificial intelligence in autonomous systems has significantly contributed to advance the research on multi-object tracking, adopted in several real-time applications (e.g., autonomous driving, surveillance drones, robotics) to localize and follow the trajectory of multiple objects moving in front of a camera. Current tracking algorithms can be divided into two main categories: some approaches introduce complex heuristics and re-identification models to improve the tracking accuracy and reduce the number of identification switches, without particular attention to the timing performance, whereas other approaches are aimed at reducing response times by removing the re-identification phase, thus penalizing the tracking accuracy. This work proposes a new approach to multi-class object tracking that allows achieving smaller and more predictable execution times, without penalizing the tracking performance. The idea is to reduce the problem of matching predictions with detections into smaller sub-problems by splitting the Hungarian matrix by class and invoking the second re-identification stage only when strictly necessary for a smaller number of elements. The proposed solution was evaluated in complex urban scenarios with several objects of different types (as cars, trucks, bikes, and pedestrians), showing the effectiveness of the multi-class approach with respect to state of the art trackers.