Multitarget tracking using mean-shift with particle filter based initialization
Satoshi Yonemoto, Motonori Sato · 2008
This paper presents a multitarget tracking algorithm based on mean-shift, in which an automatic initialization process by particle filter is utilized. In the standard mean-shift algorithm, this initialization is necessary to construct the reference target model to track. In our approach, switching between the particle filter based detection and the mean-shift tracking is introduced. Furthermore, we extend mean-shift tracking with particle filter based initialization into multitarget tracking problem. We experimentally show that our method can track multiple targets in outdoor situation and can run in real-time on a PC.