A Descriptor-Based Advanced Feature Detector for Improved Visual Tracking

Despite years of work, a robust, widely applicable generic “symmetry detector” that can paral-lel other kinds computer vision/image processing tools for the more basic structural charac-teristics, such as “edge” or “corner” detector, remains computational challenge. A new symmetry feature detector with descriptor is proposed in this paper, namely Simple Robust Features (SRF) algorithm. performance comparison made among SRF SRF, Speeded-up (SURF) SURF, Maximally Stable Extremal Regions (MSER) Harris Fast Retina Keypoint (FREAK), Minimum Eigenvalue FREAK, from Accelerated Segment Test (FAST) and Binary Invariant Scalable Keypoints (BRISK) FREAK. visual tracking dataset used evaluation terms accuracy cost. The results have shown combining preferable, it has on average highest accuracy. Additionally, cost much lower than others.