Adaptive spatio-temporal context learning for visual tracking

ABSTRACT

In recent years, a spatio-temporal context (STC) algorithm has attracted the attention of scholars, due to the algorithm makes full use of the information of the target background. Although the STC algorithm achieve tracking at the real-time, but there is still a need to improve the tracking capability when the target is occluded or the size of the target changes. In this paper, we presented an adaptive spatio-temporal context learning for visual tracking (AFSTC). Firstly, in order to accurately describe the appearance of the target, we integrate Histogram of Oriented Gradient (HOG) and Colour-naming (CN) features. And then we use the average difference between two adjacent frames to adjust the learning rate of update model for adaptive tracking. Finally, we adjust parameters of scale update strategy to achieve the competitive results on accuracy and robustness. We perform experiments on the Online Tracking Benchmark (OTB) 2015 dataset. Our tracker achieves a 13% relative gain in distance precision compared to the traditional STC algorithm. Moreover, although the speed of our tracker reduces, but it reaches 129.99 frames per second (FPS) and can still achieve tracking at the real-time.

KEYWORDS:

• Target tracking
• spatio-temporal context
• Histogram of Oriented Gradient; Colour-naming

Acknowledgement

The authors would like to thank the anonymous reviewers for their constructive comments that helped to improve the quality of this paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Yaqin Zhang (1993-) is now in the 3^rd year for her M.S in Information Science and Engineering at Xinjiang University. And her major is object tracking.

Liejun Wang (1975-) received his Ph.D. degree in the School of Information and Communication Engineering from the Xi’an Jiaotong University in 2012. He is also a member of the Education Information Teach and Teaching Committee, member of the expert group for promoting domestic cryptography in key areas in Xinjiang, director of the Advisory Committee of Educational Information Technology Experts in Xinjiang, director of the main node of the China Education and Scientific Research Network in Xinjiang, and deputy director of Network Centre. He has presided over 4 national projects related to network information security, 2 provincial and ministerial level, published more than 50 core papers. His current research interests include wireless sensor network, encryption algorithm and image intelligent processing.

Jiwei Qin (1978-) received her Master and Ph.D. degree in the School of Computer Architecture from the Xi’an Jiaotong University, Xi’an, China, in 2008 and 2013. Her research interests include intelligent network, data mining, social network modeling and analysis, E-learning, recommender systems and collaborative filtering. She has directed and participated in quite a number of research projects and published papers in many international journals and conference proceedings, including Educational Technology and Society, Sensors and Transducers, SAI Computing Conference.

ORCID

Liejun Wang http://orcid.org/0000-0002-3210-0917

Additional information

Funding

This work was supported in part by the National Science Foundation of China [under grant number 61771416] and Creative Research Groups of Higher Education Xinjiang Uygur Autonomous Region [under grant number XJEDU2017T002], the CERNET Innovation Project [under grant numbers NGII20170325 and NGII20180201].