Paper
15 September 2021 Correlation filter tracking based on multi-peak detection and adaptive coefficient
Yan-jie Zhang, Gang Wang, Yun Gao
Author Affiliations +
Proceedings Volume 11915, International Conference on Optics and Image Processing (ICOIP 2021); 119150K (2021) https://doi.org/10.1117/12.2605948
Event: International Conference on Optics and Image Processing (ICOIP 2021), 2021, Guilin, China
Abstract
In order to further improve the robustness of the ambiguity suppression related filtering algorithm (ARCF) tracking in many complex scenes such as rapid motion and occlusion, an adaptive normal punishment coefficient is proposed based on the relevant filter algorithm of automatic space-time normalization. At the same time, in order to make the tracking results more accurate, a object peak detection and selection algorithm is proposed specifically for multi-peak situation, which can determine the peak condition of the response map of the current frame and select the peak that best meets the object in multi-peak situation. In the template update stage, in order to keep the updated template from the pollution of similarities and masks, this paper also proposes a situation to detect whether the object is obscured, as a condition for the template update, so that the object template can better match the object.
© (2021) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Yan-jie Zhang, Gang Wang, and Yun Gao "Correlation filter tracking based on multi-peak detection and adaptive coefficient", Proc. SPIE 11915, International Conference on Optics and Image Processing (ICOIP 2021), 119150K (15 September 2021); https://doi.org/10.1117/12.2605948
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Detection and tracking algorithms

Electronic filtering

Image filtering

Video

Digital filtering

Automatic tracking

Optical tracking

RELATED CONTENT

A robust visual tracking via nonlocal correlation filters
Proceedings of SPIE (October 16 2019)

Back to Top