Occlusion Handling in Videos Object Tracking: A Survey

8th International Symposium of the Digital Earth

地球科学;计算机科学

Lee, B.Y.^1 ; Liew, L.H.^1 ; Cheah, W.S.^2 ; Wang, Y.C.^2

Faculty of Mathematics and Computer Sciences, Universiti Teknologi MARA, Malaysia^1

Faculty of Computer Sciences and Information Technology, Universiti Malaysia Sarawak, Malaysia^2

关键词: Articulated object;    Background scenes;    Mean-Shift tracker;    Non-linear dynamics;    Non-rigid objects;    Object to objects;    Occlusion handling;    Particle Filtering;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/18/1/012097/pdf
DOI  :  10.1088/1755-1315/18/1/012097

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

Object tracking in video has been an active research since for decades. This interest is motivated by numerous applications, such as surveillance, human-computer interaction, and sports event monitoring. Many challenges related to tracking objects still remain, this can arise due to abrupt object motion, changing appearance patterns of objects and the scene, non-rigid object structures and most significant are occlusion of tracked object be it object-to-object or object-to-scene occlusions. Generally, occlusion in object tracking occur under three situations: self-occlusion, inter-object occlusion by background scene structure. Self-occlusion occurs most frequently while tracking articulated objects when one part of the object occludes another. Inter-object occlusion occurs when two objects being tracked occlude each other whereas occlusion by the background occurs when a structure in the background occludes the tracked objects. Typically, tracking methods handle occlusion by modelling the object motion using linear and non-linear dynamic models. The derived models will be used to continuously predicting the object location when a tracked object is occluded until the object reappears. Example of these method are Kalman filtering and Particle filtering trackers. Researchers have also utilised other features to resolved occlusion, for example, silhouette projections, colour histogram and optical flow. We will present some result from a previously conducted experiment when tracking single object using Kalman filter, Particle filter and Mean Shift trackers under various occlusion situation in this paper. We will also review various other occlusion handling methods that involved using multiple cameras. In a nutshell, the goal of this paper is to discuss in detail the problem of occlusion in object tracking and review the state of the art occlusion handling methods, classify them into different categories, and identify new trends. Moreover, we discuss the important issues related to occlusion handling including the use of appropriate selection of motion models, image features and use of multiple cameras.