Feature-Scoring-Based Multi-cue Infrared Object Tracking
In this paper, we propose an effective tracker for infrared videos based on the multi-cue fusion. Under the particle filter tracking construction, a novel feature scoring scheme is introduced to evaluate different cue tracking ability, then the multi-cue fusion is executed in a weighted sum manner. In our tracking system, the score of each feature can be adaptively updated according to the current environment. Experimental results with various Infrared Video Database and different trackers are reported to demonstrate the accuracy and robustness of our algorithm.
Keywordsinfrared video object tracking multi-cue fusion particle filter
Unable to display preview. Download preview PDF.
- 1.Wang, R., Popovic, J.: Real-Time Hand-Tracking with a Color Glove. ACM Trans. Graph. 28(3) (2009)Google Scholar
- 6.OTCBVS benchmark dataset, http://www.cse.ohio-state.edu/otcbvs-bench/
- 7.VIVID database, https://www.sdms.afrl.af.mil/request/data-request.php#vivid