Probabilistic Change Detector with Human Verifier for Intelligent Sterile Zone Monitoring
Ajmal Shahbaz, Kang-Hyun Jo · 2018
This paper proposed change detector algorithm with human verifier for intelligent surveillance systems in general and sterile zone monitoring in particular. The proposed algorithm can be used during day and night time. It is designed to tackle practical problems such as illumination change, static foreground object, and camouflage foreground object. It uses statistical criteria which helps to detect switch between the color and IR camera. The proposed algorithm uses change detection using Gaussian Mixture Model (GMM) as region proposal. It gives possible region of interest to be intruder. Later, that RP is fed into human verifier which is HOG based SVM classifier. It labels intruder if it is human. The proposed method is tested on the standard datasets.