Update: This story has been revised throughout to clarify the use of Briefcam in the investigation. 7/28/13 18:55 CST.
aNewDomain – As the deployment of video surveillance cameras grows, efficient and effective viewing and analysis of video streams becomes costly and labor intensive task. The monotonous task of monitoring surveillance video leads to fatigue, and “not knowing what one is even looking for” can render this powerful tool ineffective.
The surveillance challenge – as Amit Gavish of BriefCam Americas puts it – is an issue of the huge investment of video surveillance equipment traditionally returning a very low percentage of utilization. While 100 percent of video is recorded 24/7, only 8 percent of that video is “viewed”, and less than 2 percent is “reviewed” by investigators.
Therefore 90 percent of the video is never “viewed” and 98 percent is never “reviewed.”
While recent revelations of alleged widespread surveillance and potential breaches of personal privacy have caused public outcry, we have also seen how widespread surveillance led to the apprehension of the suspects in the Boston Marathon bombings. However, according to Israeli news sources, investigators had a powerful new tool to assist them with viewing the huge volume of both pre-recorded and live video streams, contributing to the identification and apprehension.
Israel-based BriefCam, with its U.S. office located in Farmington, CN develops “Video Synopsis” software enabling the simultaneous presentation of “events” which occur at different times in a video stream. The solution enables security personnel to implement on-the-spot action and provides investigators with a faster solution for post-event investigations.
Originally developed by Professor Shmuel Peleg at The Hebrew University of Jerusalem, Israel, Video Synopsis software summarizes streams of video by tracking and analyzing moving objects (also called “events” or “activities”) and translates these video streams into a database of “objects” “events” and “activities”.
The software is format agnostic and assists users in efficiently reviewing lengthy video streams, rather than take the human element out of the loop all together. This creates a good balance between the efficiency of software processing and human discretion.
This process begins with the “Ingestion” stage, where the video stream is analyzed for static, non-moving objects – such as the background – and dynamic objects which change over time. It’s the dynamic objects that are of interest; these are extracted from the frames of video and automatically entered into a database.
In the next stage, a “Synopsis” is automatically generated based on a user specified time period of interest (e.g. the hours preceding an event), where multiple “dynamic objects” are shifted in time and shown simultaneously, even though they appeared at different times. The result is a much shorter video that can often summarize a full day of video into a few minutes of video.
In the final stage the “dynamic objects” are indexed according to where they occurred in the original video stream, enabling the user to select an object of interest and jump to the segment of video where the object appeared.
Once the events of interest are located, the original footage can then be analyzed using facial recognition, heat mapping and other software to further evaluate what’s contained in the video. BriefCam leaves this analysis to other software companies, focusing on the processing of hours of video streams into minutes of Video Synopsis.
BriefCam eases privacy concerns because the video being processed is recorded and monitored anyway, and BriefCam’s software limits the amount and scope of what’s being reviewed.Tags: Security,Technology