Introduction: Signal sparse representation is a very interesting field of research in the signal processing community in the past 20 years. Many research papers and symposiums have shown that the field is booming. The purpose of signal sparse representation is to represent the signal with as few atoms as possible in a given overcomplete dictionary and to obtain a more concise representation of the signal, thus making it easier for us to obtain the information contained in the signal. Further processing of signals, such as compression, coding, etc. This article describes a new approach to SSVT that can more effectively implement visual tracking.
Set to Set Visual Tracking
Abstract: Sparse representation has been widely used in visual tracking, achieving excellent tracking results. However, most sparse representation models represent candidate targets as a linear combination of target templates, and sparse optimization problems need to be solved. In this paper, we propose a new set-to-set visual tracking (SSVT) method. Under the framework of particle filter, the candidate targets and target templates are considered as image sets, and they are modeled as convex hulls. Then the distance between the two image sets is minimized, and the tracking result is a candidate target with the largest coefficient. When candidate targets are modeled as convex hulls, SSVT exploits the potential relationship between candidate targets. In addition, SSVT works well in this area. It only needs to solve a second-order optimization problem, instead of solving the sparse optimization problem. Qualitative and quantitative analysis results on several challenging image sequences show that the SSVT algorithm proposed in this paper is superior to the best performance tracker.
Keywords: set-to-set distance, visual tracking, particle filter, convex hull, support vector machine
Via:PRICAI 2016
PS : This article was compiled by Lei Feng Network (search "Lei Feng Network" public number) and it was compiled without permission.
Original paper download
