Intelligent video anomaly detection and classification using faster RCNN with deep reinforcement learning model

Recently, intelligent video surveillance applications have become essential in public security by the use of computer vision technologies to investigate and understand long streams. Anomaly detection classification are considered a major element surveillance. The aim anomaly is automatically determine existence abnormalities short time period. Deep reinforcement learning (DRL) techniques can be employed for detection, which integrates concepts deep enabling artificial agents knowledge experience from actual data directly. With this motivation, paper presents an Intelligent Video Detection Classification using Faster RCNN with Reinforcement Learning Model, called IVADC-FDRL model. presented model operates on two stages namely classification. Firstly, applied as object detector Residual Network baseline model, detects anomalies objects. Besides, Q-learning (DQL) based DRL detected anomalies. In order validate effective performance extensive set experimentations were carried out benchmark UCSD dataset. experimental results showcased better over other compared methods maximum accuracy 98.50% 94.80% Test004 Test007 dataset respectively.