A fast on-line boosting tracking algorithm based on cascade filter of multi-features
- DOI
- 10.2991/icmt-13.2013.226How to use a DOI?
- Keywords
- object tracking; on-line boosting; cascade detection; OSS model; color histogram
- Abstract
Exhaustive search in local area around the object is used for tracking in the conventional Haar-like feature based on-line boosting tracking methods (HBT). Tracking is frequently failed and can hardly find the object again when the object moves fast or occluded completely. In addition, on-line tracking will generate cumulative errors during updating when the state of the object and the environment changes. In this paper, we propose a fast on-line boosting tracking algorithm based on cascade filters of multi-features. Algorithm generates detection windows with various positions and sizes during initialization. Cascade detection screenings possible object windows by using multi-features: the manner of object motion, variance of detection windows, Object Similarity Statistic (OSS) model, color distribution and overlap ratio in turn. The left candidate object windows are passed to the strong classifier of on-line boosting and the object is located finally. In order to exert the advantage of boosting, naive Bayes classifier is used to replace the original simple threshold classifier in HBT. Experiment results on various videos show that the robustness, accuracy and real-time ability are improved.
- Copyright
- © 2013, the Authors. Published by Atlantis Press.
- Open Access
- This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Cite this article
TY - CONF AU - Song Hu AU - Shui-Fa Sun AU - Xian-Bing Ma AU - Yin-Shi Qin AU - Bang-Jun Lei PY - 2013/11 DA - 2013/11 TI - A fast on-line boosting tracking algorithm based on cascade filter of multi-features BT - Proceedings of 3rd International Conference on Multimedia Technology(ICMT-13) PB - Atlantis Press SP - 1863 EP - 1870 SN - 1951-6851 UR - https://doi.org/10.2991/icmt-13.2013.226 DO - 10.2991/icmt-13.2013.226 ID - Hu2013/11 ER -