Adaptive Mean-shift Video Object Tracking with Color and Depth Information
碩士 === 義守大學 === 資訊工程學系碩士班 === 99 === Mean-shift is a well-known approach for tracking objects in video sequences. However, it suffers from the problems of limited discriminative ability of color features and static target model in the whole tracking process. Therefore, we propose an adaptive Mean-sh...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2011
|
Online Access: | http://ndltd.ncl.edu.tw/handle/58649428731039194006 |
Summary: | 碩士 === 義守大學 === 資訊工程學系碩士班 === 99 === Mean-shift is a well-known approach for tracking objects in video sequences. However, it suffers from the problems of limited discriminative ability of color features and static target model in the whole tracking process. Therefore, we propose an adaptive Mean-shift tracking approach with color and depth information for solving these problems. The depth feature is combined with color features for modeling the target object and target candidates. Besides, an adaptive learning is employed to update the target model in the tracking process. Experimental results have shown that our approach achieves better tracking results than traditional mean-shift.
|
---|