Title :
Three dimensional trajectories mining
Author :
Tsumoto, Shusaku ; Hirano, Shoji
Author_Institution :
Dept. of Med. Inf., Shimane Univ., Izumo, Japan
Abstract :
In this paper, we propose a multiscale comparison method for three-dimensional trajectories based on the maxima on the curvature scale space. We define a segment as a partial trajectory between two adjacent maxima where curvature becomes locally maximal. Then we trace the place of maxima across the scales in order to obtain the hierarchy of segments. By applying segment-based matching technique, we obtain the best correspondences between partial trajectories. We demonstrate in a preliminary experiment that our method could successfully capture structural similarity in three-dimensional trajectories.
Keywords :
computational geometry; data mining; medical information systems; pattern clustering; pattern matching; curvature scale; data clustering; segment based matching; temporal data mining; three dimensional trajectory mining; Matched filters; Trajectory; Clustering; Multiscale comparison; Temporal data mining; Trajectories;
Conference_Titel :
Systems Man and Cybernetics (SMC), 2010 IEEE International Conference on
Conference_Location :
Istanbul
Print_ISBN :
978-1-4244-6586-6
DOI :
10.1109/ICSMC.2010.5641969