Abstract
We propose a high-performance visual target tracking (VTT) algorithm based on classified-patch kernel particle filter (CKPF). Novel features of this VTT algorithm include sparse representations of the target template using the label-consistent K-singular value decomposition (LC-KSVD) algorithm; Gaussian kernel density particle filter to facilitate candidate template generation and likelihood matching score evaluation; and an occlusion detection method using sparse coefficient histogram (ASCH). Experimental results validate superior performance of the proposed tracking algorithm over state-of-the-art visual target tracking algorithms in scenarios that include occlusion, background clutter, illumination change, target rotation, and scale changes.
Highlights
Visual target tracking (VTT) [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15] is a key enabling technology for numerous emerging computer vision applications including video surveillance, navigation, human-computer interactions, augmented reality, higher level scene understanding, and action recognition among many others
In this work, we propose an adaptive visual target tracking algorithm based on classified-patch kernel particle filter (CKPF), which has the following advantages: (a) Classified patches and low-dimensional dictionary are considered in the CKPF
L1 tracker using accelerated proximal gradient (L1APG), Multi-task tracking (MTT), and Incremental learning visual tracking (IVT) cannot extract the target correctly due to the use of the fixed global model, while the proposed algorithm employs the local patch features to describe the details of the target, and the labelconsistent K-singular value decomposition (LC-KSVD) method is introduced to learn dictionaries and train the classification parameters simultaneously, which can decrease the influence of the background disturbance
Summary
Visual target tracking (VTT) [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15] is a key enabling technology for numerous emerging computer vision applications including video surveillance, navigation, human-computer interactions, augmented reality, higher level scene understanding, and action recognition among many others It is a challenging task because the visual observations often suffer from interference due to occlusion, scale and shape variation, illumination variation, background clutter, and related factors. In the 434th frame of the Girl sequences, the face of the girl is occluded by the man and the scale makes a little change during the process of target rotation; the proposed algorithm obtains a good tracking result. From the Board sequences, we can draw the same conclusions, in which the proposed algorithm has a good performance of target tracking under the scenario with target rotation and scale variation
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
