The FINSEC CCTV analytics probe consists of a probe producing events coming from observations of physical interactions by CCTV and provides physical threat intelligence to the FINSEC cyber-physical threat intelligence framework. The goal is to be able to detect objects (such as bodies, heads, hands, etc.) and capture their interactions with each other and with physical motionless objects (e.g. ATMs or data center racks). These observations generate events that are pushed to the FINSEC data collection for further analytics and correlation.
When it comes to privacy CCTV analytics is a sensitive field: especially automated analysis and event generation require scrutinized attention. Within FINSEC Fujitsu develops an AI based technology, which enables automation of CCTV analytics fulfilling GDPR requirements in the same time.
The Analytics probe scans the video stream, detecting bodies or body parts in marked areas of a scene. Then detected bodies are tracked in the scene and poses and actions are analysed. Based on this, anonymous events such as a body entering or leaving a marked zone, are generated. These are sent to the FINSEC platform using the FINSTIX model. Finally, the FINSEC platform utilizes these physical events for further security analysis and cyber physical threat correlation.
A second version of the FINSEC CCTV Analytics Probe is currently under development; in thsi new version the limitations of the first implemented probe shall be reduced and the detection enhanced. Moreover, tracking of bodies shall be improved: for example, the evaluation of velocity of bodies shall improve detection of bodies and reduce the limitations due to the occlusion of moving bodies by each other. This also will enable the generation of new events such as people moving with high velocity.