Single target tracking via correlation filter and context adaptively
- PDF / 2,143,636 Bytes
- 18 Pages / 439.37 x 666.142 pts Page_size
- 77 Downloads / 172 Views
Single target tracking via correlation filter and context adaptively Hua Bao 1 & Yixiang Lu 1 & Qijun Wang 1 Received: 29 August 2019 / Revised: 8 June 2020 / Accepted: 9 July 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
Recently, tracking methods based on correlation filter show impressive performance in a variety of complex environments for their excellent classification performance. However, most of the existing methods only focus on the information in the bounding box regions of the target, and do not fully utilized the contextual information and the internal structure information of the target. Thus, when the target scenes experience dramatic change, the learned filter could not accurately adapt to the appearance change, which will lead to model degradation of the target. To address this issue, a novel approach via correlation filter and context jointly is proposed in this paper. First, we decompose the target into multiple independent parts, and each part learns the filtering response separately. Through the joint learning of multiple independent filters, the target model can effectively maintain the structural information of the object and is not sensitive to partial occlusion, and etc. Second, we introduce multi-channel features in the representation based on the parts of the target and the contextual information to migrate the background influences. With the introduction of collaborative representation strategy, the impact of background noise can be effectively suppressed. To evaluate the proposed approach, we conduct extensive experiments on several challenging benchmark datasets including OTB-2013 and OTB2015 datasets. The results show our method demonstrates comparable performance against several state-of-the-art methods. Keywords Visual tracking . Correlation filter . Contextual information
* Qijun Wang [email protected]
1
School of Electrical Engineering and Automation, Anhui University, Hefei 230601, People’s Republic of China
Multimedia Tools and Applications
1 Introduction Visual tracking is one of hotspots in the field of computer vision, and its purpose is to accurately estimate the state of a pre-specified target including positions and scales in continuous image sequences. It has a variety of applications in video security, public surveillance, robot navigation, to name a few. Although significant performance improvements have been achieved in the past few decades. However, how to attain robust performance remains a challenge, mainly due to significant appearance variations of the target caused by illumination, pose, occlusion, and so on [9, 18, 22, 36]. Therefore, how to learn an effective target model, which could handle different challenging variations is one of the key issues in this topic. Recently, tracking approaches based on tracking-by-detection scheme have received significant attention for their effectiveness and robustness [19, 21, 25]. These approaches predict the state of the target by learning an accurately classification model, wh
Data Loading...
