Modeling local and global behavior for trajectory classification using graph based algorithm

• Published in: Pattern recognition letters Vol. 150; pp. 280 - 288
• Main Authors: Saini, Rajkumar, Kumar, Pradeep, Roy, Partha Pratim, Pal, Umapada
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.10.2021; Elsevier Science Ltd
• Subjects: Algorithms; Classification; Clusters; Complete bipartite graph; Dynamic time warping (DTW); Graph theory; Minimum spanning tree (MST); Particle swarm optimization; Particle swarm optimization (PSO); Similarity; Trajectory classification and clustering; Vigilance; Particle swarm optimization (PSO); Complete bipartite graph; Minimum spanning tree (MST); Trajectory classification and clustering; Dynamic time warping (DTW)
• ISSN: 0167-8655; 1872-7344
• DOI: 10.1016/j.patrec.2019.05.014

•A graph based trajectory classification using local & global behavior is proposed.•Global behavior is computed from full length trajectories.•Complete Bipartite Graph is used for modeling local behavior from trajectory slices.•Global and local behaviors are combined using particle swarm optimization.•Proposed method has been tested on public datasets. Understanding motion patterns is of great importance to analyze the behavior of objects in the vigilance area. Grouping the motion patterns into clusters in such a way that similar motion patterns lie in same cluster and the inter-cluster variance is maximized in a challenging work. Variation in the duration of trajectory patterns in terms of time or number of points in them (even in the trajectories from same cluster) makes it more difficult to correctly classify in respective clusters while using full length trajectories, local clue can be used along with the global information. Trajectories can be segmented into distinctive parts and local contribution of these parts can be used to improve the performance of the system. In this work, we have formulated the trajectory classification problem into graph based similarity problem using Douglas–Peucker (DP) algorithm, Complete Bipartite Graphs (CBG), and Minimum Spanning Tree (MST). Local behavior of objects has been analyzed using their motion segments and Dynamic Time Warping (DTW) has been used for finding similarity among motion trajectories. Class-wise global and local costs have been computed using DTW, CBG, and MST and their fusion has been done using Particle Swarm Optimization (PSO) to improve the classification rate. Trajectory datasets, namely T15, LabOmni, and CROSS have been used in experiments. The proposed method yields encouraging results and outperforms the state of the art techniques.