Acoustic and visual signal based violence detection system for indoor security application

We propose an acoustic and visual signal based person-to-person violence detection system for indoor security application. A final decision is determined by fusing those detection results from each visual and acoustic detector. Information fusion is made with a Bayesian framework.