Low-cost multiple object tracking for embedded vision applications

Abstract

This paper presents a low-cost multiple object tracking (MOT) technique by employing a novel appearanceupdate model for object appearance modeling using K-means. The state-of-the-art work has attained a very highaccuracy without considering the real-time aspects necessitated by currently trending embedded vision platforms. Themajor research on multiple object tracking is used to update the appearance model in every frame while discounting itspersistent nature. The proposed appearance update model reduces the computational cost of the state-of-the-art MOT6-fold by exploiting this facet of persistent appearance over the sequence of frames. To ensure accuracy, the proposedmodel is tested on different publicly available standard datasets with challenging situations for both indoor and outdoorscenarios. The experimental results illustrate that our model successfully achieves multiple object tracking while copingwith long-term and complete occlusion. The proposed method achieves the same accuracy in comparison with thestate-of-the-art baseline methods. Moreover, and most importantly, the proposed method is cost-effective in terms ofcomputing and/or memory requirements in comparison to the state-of-the-art techniques. All these traits make ourdesign very suitable for real-time and embedded video surveillance applications with low computing/memory resources.