Housebreaking crime gait pattern classification using artificial neural network and support vector machine

• Published in: Journal of Theoretical and Applied Information Technology
• Main Author: Razak H.A.; Almisreb A.A.; Mohammed Saleh M.A.; Tahir N.M.
• Format: Article
• Language: English
• Published: Little Lion Scientific 2020
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85090516529&partnerID=40&md5=13884737cd33a0a6ea6cbc12aa349b7b

The rate of crime is worsen and has led to a growing number of studies on human identification namely gait recognition. Hence, this study focused on the normal and anomalous behavior at the gate of residential units based on gait features extracted using Kinect sensor. Firstly, dataset of housebreaking crime behavior and normal behavior at the gate is acquired and collected. Further, orthogonal least squares (OLS) are utilized to extract and select the gait features along with principal component analysis (PCA) as gait feature optimization. Next, classification of gait features is done using artificial neural network (ANN) and support vector machine (SVM). Result attained showed that the recognition performance using ANN classifier was up to 99% but only 50% for SVM classifier. Findings from this study showed that the most optimum accuracy rate is at 99.78% using ANN with GDX as the learning algorithm in classifying both normal and anomalous behavior at the residential gate units. © 2020 Little Lion Scientific. All rights reserved.