Abstract
In this paper, we propose a biologically inspired appearance model for robust visual tracking. Motivated in part by the success of the hierarchical organization of the primary visual cortex (area V1), we establish an architecture consisting of five layers: whitening, rectification, normalization, coding, and pooling. The first three layers stem from the models developed for object recognition. In this paper, our attention focuses on the coding and pooling layers. In particular, we use a discriminative sparse coding method in the coding layer along with spatial pyramid representation in the pooling layer, which makes it easier to distinguish the target to be tracked from its background in the presence of appearance variations. An extensive experimental study shows that the proposed method has higher tracking accuracy than several state-of-the-art trackers.
Highlights
V ISUAL tracking is a task that continuously infers the state of a specific target from an image sequence; it is a specific task of computer vision which has attracted increasing interest in recent years [1]–[15]
In this work, we propose a biologically inspired appearance model for robust visual tracking
In visual tracking applications, we need to maintain the discriminative capability of the appearance model employed in the tracker all the time, which is more challenging than the case of object recognition, and our attention in this paper focuses on the coding and pooling layers
Summary
V ISUAL tracking is a task that continuously infers the state of a specific target from an image sequence; it is a specific task of computer vision which has attracted increasing interest in recent years [1]–[15]. 4) Template updating, where we update the appearance of the target template in order to adapt to the variations of the target appearance over time In these four stages, appearance modeling attracts large attention in the last few decades [3], [16]–[29]. Several recent models resort to machine learning methods and devise suitable training criteria and optimization methods to learn features from data [33]– [35]. In spite of their high adaptive capability, these methods still somehow depend on prior knowledge, which does not transfer to other applications
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 callMore From: IEEE Transactions on Neural Networks and Learning Systems
Disclaimer: 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.
