The kernel based tracking has two disadvantages: the tracking window size cannot be adjusted efficiently, and the kernel based color distribution may not have enough ability to discriminate object from clutter background. For boosting up the feature's discriminating ability, both scale invariant features and kernel based color distribution features are used as descriptors of tracked object. The proposed algorithm can keep tracking object of varying scales even when the surrounding background is similar to the object's appearance.
© 2008 Chinese Optics Letters
(070.5010) Fourier optics and signal processing : Pattern recognition
(100.0100) Image processing : Image processing
(330.0330) Vision, color, and visual optics : Vision, color, and visual optics
(330.7310) Vision, color, and visual optics : Vision
Risheng Han, Zhongliang Jing, and Yuanxiang Li, "Kernel based visual tracking with scale invariant features," Chin. Opt. Lett. 6, 168-171 (2008)