Untitled Document
Untitled Document
> Archives > Current Issues
JEET, Vol. 13, No. 2, March 2018
Spatio-Temporal Analysis of Trajectory for Pedestrian Activity Recognition
Moon-Hyun Kim
Area D - Information and Control
Abstract Recently, researches on automatic recognition of human activities have been actively carried out with the emergence of various intelligent systems. Since a large amount of visual data can be secured through Closed Circuit Television, it is required to recognize human behavior in a dynamic situation rather than a static situation. In this paper, we propose new intelligent human activity recognition model using the trajectory information extracted from the video sequence. The proposed model consists of three steps: segmentation and partitioning of trajectory step, feature extraction step, and behavioral learning step. First, the entire trajectory is fuzzy partitioned according to the motion characteristics, and then temporal features and spatial features are extracted. Using the extracted features, four pedestrian behaviors were modeled by decision tree learning algorithm and performance evaluation was performed. The experiments in this paper were conducted using Caviar data sets. Experimental results show that trajectory provides good activity recognition accuracy by extracting instantaneous property and distinctive regional property.
Keyword Decision tree,Fuzzy partitioning,Human activity recognition,Spatial feature,Trajectory 1. Introduction Recognizing human activities is at the heart of user interfaces and applications for smart environments. It is also applied to various technologies such as surveillance security,human-computer interaction,content-based video retrieval,and virtual reality. In particular,automatic recognition of human activities is one of the most important research tasks in the field of machine learning and computer vision. And visual data is one of the most important clues in the development system for understanding and applying human behavior correctly [1,2]. Automatic recognition of human activities requires two processes [3]: i) extraction of activity information,and ii) activity pattern modeling. Activity information refers to the movement attributes (speed,direction,location,etc.) of data and it is necessary to extract effective features for classifying human activities showing various behavior patterns. Activity patterns are expressions of frequently occurring events [4] and should be accompanied by appropriate machine learning algorithms to learn human activity patterns. Naftel et al. [5] introduced a classification methodology for anomaly detection. His theory is related with the utilization of unsupervised learning in coefficient feature space for spatiotemporal objects trajectories. Learning trajectory patterns with Self-Organizing Maps well classifies abnormality from trajectories. But,it only detects in
Untitled Document