Multiple Object Tracking through Background Learning Deependra Sharma * and Zainul Abdin Jaffery Jamia Millia Islamia, Department of Electrical Engineering, New-Delhi, 110025, India *Corresponding Author: Deependra Sharma. Email: medeependrasharma@gmail.com Received: 18 September 2021; Accepted: 14 December 2021 Abstract: This paper discusses about the new approach of multiple object track- ing relative to background information. The concept of multiple object tracking through background learning is based upon the theory of relativity, that involves a frame of reference in spatial domain to localize and/or track any object. The field of multiple object tracking has seen a lot of research, but researchers have considered the background as redundant. However, in object tracking, the back- ground plays a vital role and leads to definite improvement in the overall process of tracking. In the present work an algorithm is proposed for the multiple object tracking through background learning. The learning framework is based on graph embedding approach for localizing multiple objects. The graph utilizes the inher- ent capabilities of depth modelling that assist in prior to track occlusion avoidance among multiple objects. The proposed algorithm has been compared with the recent work available in literature on numerous performance evaluation measures. It is observed that our proposed algorithm gives better performance. Keywords: Object tracking; image processing; background learning; graph embedding algorithm; computer vision 1 Introduction Multiple object tracking (MOT) is a crucial task in computer vision, with a wide range of tracking algorithms. Tracking and its complexity level depends on several factors, such as type of parameters being tracked namely size, contour, position, velocity, and acceleration. It may also depend on number of parameters used for tracking and the amount of prior knowledge about the target object. During tracking different situations may arise such as, tracking of mobile object appearing for the first time in the scene. When representations of the object under consideration are available, it is feasible to learn it for the first time. Object tracking is an act of seeking for objects in successive frames of a video stream after learning has been completed. Even after so much research, MOT remains a difficult work since the object’ s appearance can radically vary due to deformation, rotation out of plane, or changes in lighting conditions. Problem becomes more challenging when tracking is to be done in dense places that consists of movable and immovable objects. When faced handling problems including occlusions, illumination changes, motion blur, and other environmental changes, most existing techniques underperform [1]. To solve these issues, we present a background learning-based multiple object tracker. Even if background removal results in the identified This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Computer Systems Science & Engineering DOI: 10.32604/csse.2023.023728 Article ech T Press Science