A Study on Abnormal Behavior Detection in CCTV Images through the Supervised Learning Model of Deep Learning
Journal of Logistics Informatics and Service Science · 2022
Abnormal behavior refers to behavior that is atypical or statistically uncommon within a particular culture or that is maladaptive or detrimental to an individual or to those around that individual.Most of these actions are criminal acts that people fear.The installation of public CCTVs to prevent these crimes is increasing, but the crime rate is rather increasing recently.In line with this situation, artificial intelligence research using deep learning that automatically finds abnormal behavior in CCTV is increasing.Deep learning is a type of artificial intelligence designed based on artificial neural networks.C3D is a model that uses 3D Convolution Networks to train large amounts of image data in a supervised learning manner.In this paper, a study was conducted to learn and detect abnormal behavior images consisting of five crime types and 32 detailed types using the C3D model.As a result of learning and classifying abnormal behavior images with the C3D model, which is an artificial intelligence model, abnormal behavior was detected in several crime type image data, showing considerably high accuracy.Therefore, it is believed that abnormal behavior detection technology using C3D can contribute to protecting victims by classifying the precursor symptoms of crime by applying it to various CCTV image data to prevent crime in advance or detect criminal behavior.