Paper on Multi-Object Tracking in Videos published on IEEE Access

Our paper titled “Efficient Online Tracking-by-Detection With Kalman Filter” has been published on IEEE Access. The paper was co-authored by Siyuan Chen and Prof. Chenhui Shao.

Visual Multi-Object Tracking (MOT) has a promisingly broad application in manufacturing, construction, traffic, logistics, etc., especially in large-scale applications where it is not feasible to attach markers to many objects for traditional, marker-enabled tracking methods. This paper presents a lightweight, open-source MOT algorithm, Kalman-intersection-over-union (KIOU) tracker, for fast multi-object tracking in videos that integrates a Kalman filter with IOU-based track association methods. The KIOU tracker is quantitatively evaluated with UA-DETRAC, a dense traffic video dataset to demonstrate its performance and speed.

Read the full paper here and the source code here.