Development of CCTV-camera-based System for Detection of Anomalous Behaviors in Penitentiary Institutions
Piotr Bilski, A. Buchowicz, Brunon Hołyst, Konrad Jędrzejewski, Marcin Lewandowski, Paweł Mazurek, Jacek Olejnik · 2021 11th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS) · 2021
This paper presents the interdisciplinary project aimed at detecting the undesired or dangerous behavior of persons in the confined institutions such as prisons or wards, based on the video streams provided by the CCTV cameras. Currently, there are IT systems working in such areas, but their efficiency is limited, forcing the operator to focus his/her attention on every image separately. Design of a system for autonomous detection of anomalous behaviors - such as fights or passing illegal material - based on the input from multiple surveillance cameras would allow for providing the decision support module. With the acceptable detection accuracy, the solution would help in minimizing the accidents and unwanted behavior of monitored inmates. The system should also provide the catalogue of predefined dangerous situations, prepared by the experts in criminal sciences, including the features useful for predicting the specific events.