Raw Trajectory Rectification via Scene-Free Splitting and Stitching  

作　　者：郭春超 胡晓军 赖剑煌 石世昌 陈世哲 

机构地区：[1]School of Information Science and Technology, Sun Yat-sen University, Guangzhou 510006, China

出　　处：《Journal of Computer Science & Technology》2015年第2期364-372,共9页计算机科学技术学报（英文版）

基　　金：This work was supported by the National Science ＆ Technology Pillar Program of China under Grant No. 2012BAK16B06 and the National Natural Science Foundation of China under Grant No. 61173084.

摘　　要：Trajectories carry rich motion cues and thus have been leveraged to many high-level computer vision tasks. Due to the easy implementation of simple trackers, most previous work on trajectory-based applications utilizes raw tracking outputs without explicitly considering tracking errors. Reliable trajectories are prerequisite for modeling and recognizing high-level behaviors. Therefore, this paper tackles such problems by rectifying raw trajectories, which aims to post-process existing trajectories. Our approach firstly splits them into short tracks, and then infers identity ambiguity to remove unqualified detection responses. At last, short tracks are stitched via maximum bipartite graph matching. This post- processing is completely scene-free. Results of trajectory rectification and their benefits are both evaluated on two challenging datasets. Results demonstrate that rectified trajectories are conducive to high-level tasks and the proposed approach is also competitive with state-of-the-art multi-target tracking methods.Trajectories carry rich motion cues and thus have been leveraged to many high-level computer vision tasks. Due to the easy implementation of simple trackers, most previous work on trajectory-based applications utilizes raw tracking outputs without explicitly considering tracking errors. Reliable trajectories are prerequisite for modeling and recognizing high-level behaviors. Therefore, this paper tackles such problems by rectifying raw trajectories, which aims to post-process existing trajectories. Our approach firstly splits them into short tracks, and then infers identity ambiguity to remove unqualified detection responses. At last, short tracks are stitched via maximum bipartite graph matching. This post- processing is completely scene-free. Results of trajectory rectification and their benefits are both evaluated on two challenging datasets. Results demonstrate that rectified trajectories are conducive to high-level tasks and the proposed approach is also competitive with state-of-the-art multi-target tracking methods.

关 键 词：raw trajectory rectification trajectory post-processing identity ambiguity multi-target tracking activity classification 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TU998.12[自动化与计算机技术—计算机科学与技术]