How to quickly transform hours of video surveillance into actionable intelligence
The BriefCam Video Content Analytics Platform uses artificial intelligence to accelerate investigations, derive actionable insights and attain situational awareness from video content
Sponsored by BriefCam
By James Careless for Police1 BrandFocus
There are more than 350 million CCTV surveillance cameras in operation globally, recording petabytes of potentially actionable information for apprehending criminals and solving crimes.
For police, the challenge is to sort through this mass of video quickly and accurately. Unfortunately, this is hard to achieve using human-based video analysis. People are not equipped to retain focus when watching hours of surveillance video. They get distracted, tired, and require breaks. This means an hour of video can take three hours to be adequately analyzed by a human operator.
But what if this surveillance video could be filtered by a video content analytics platform; a software package that can allow human operators to specify what they’re looking for and filter out everything else? And what if it was able to compress footage that matched the human-defined criteria into just a few minutes of video that could be quickly reviewed?
Say a vehicle of interest is a blue four-door sedan. Enter this into the filtering software, and all objects lacking blue sedans are now eliminated from the video made available to the human operator.
Suddenly hours of surveillance video to be reviewed turns into minutes or even seconds. The human’s attention capacity is capable of working with this length of meaningful surveillance video quickly and deriving conclusions that are reliable and actionable.
The same is true for seeking human perpetrators captured by surveillance cameras. If the officer viewing the surveillance video knows the appearance, gender, and clothing of the suspect, this can be entered into the video content analytics platform; all others will be filtered out.
If a digital photo of the suspect is on file and the analytics platform incorporates facial recognition, the video can be filtered to show where this person was (or is). The system can also switch from one camera view to another; creating a timeline-based map of their movements.
This tracking technology was used to find two brothers involved in the allegedly staged attack on ‘Empire’ actor Jussie Smollett in 2019. The Chicago Police Department analyzed recorded video captured by more than four dozen CCTV cameras to track the brothers before and after the alleged attack, to determine where they lived and who they were.
The analytical platform that helped crack the Smollett case is called BriefCam. And it is every bit as powerful as the Chicago PD case suggests – and more so.
BriefCam is a video content analytics platform that connects to a broad selection of video management systems and can also be used to process pre-recorded video. Using filters selected by the human operator in an intuitive, easy-to-use interface, BriefCam digitally eliminates all frames that do not meet the selected criteria. The result is video that shows only those items that fit the criteria (eg., blue sedans) within whatever geographical area of camera coverage and timeline the operator selects.
Once this video has been filtered, the finished product can be slowed, sped up, reversed, and stopped as much as the operator desires. In a three-minute video clip of blue cars at a single intersection, each one can be selected and examined; and even cross-referenced from shots of the same vehicle at nearby intersections. All other vehicles that don’t match the search criteria simply disappear from the Video Synopsis.
The impact on investigations – especially child abductions, shootings and terrorist attacks – can be profound. It is now possible to review surveillance video hours and even days before and after an incident took place. And when facial recognition is employed, a perpetrator’s movement and actions can be traced back hours before and after an event; especially if the investigating officers know the suspect’s preferred neighborhoods and hangouts. In the same way, accessories to the suspect’s actions can also be identified.
BriefCam’s facial recognition and behavioral analysis can also help boost situational awareness in real-time; aiding police in apprehending suspects during the commission of a crime.
The power of this technology was recently proven in Hartford, Connecticut, where BriefCam helped police nab a suspected child predator quickly and safely.
BriefCam Helps Bust a Suspected Child Predator
In the summer and fall of 2017, a suspected kidnapper driving a white van was working the streets of Hartford, Connecticut, trying to lure young girls into his vehicle.
Usually, a predator like this is extremely hard to catch. Just because he drove past some of the city’s 545 surveillance cameras, doesn’t mean this 380-person Hartford Police Department (HPD) police force has the capabilities to find him on their extensive surveillance video files.
Having installed BriefCam in their department, the Hartford PD was able to use the description of the suspect’s white van and location of an attempted child luring to identify and arrest 34-year-old Jose Jalpa-Yane, a father of three. Thanks to BriefCam, the Hartford PD was able to identify the suspect’s license plate and also spot footage of the van circling the same location a few days prior.
“This would not have happened prior to having this technology,” said Hartford police Sgt. Johnmichael O’Hare, who worked with BriefCam during this case to apprehend Jalpa-Yane. “Nobody is going to be able to sit there and watch two hours of video in a few minutes and capture what they need in 13 seconds”– which is what BriefCam did for Hartford PD.
Other BriefCam Benefits
Speeding up current investigations and making cold cases solvable are just two of BriefCam’s many benefits. The analytical data compiled by BriefCam Research can be used to map out crime patterns within a given locale over time so that better policing strategies can be designed and implemented.
The BriefCam Respond module can trigger real-time alerts of wanted suspects turning up on camera when their facial recognition data is input into the system for ongoing monitoring. Potentially threatening changes in an environment (such as the build-up of crowds) can be signaled as well, multiplying a department’s ability to maintain order and public safety with limited resources.
BriefCam is platform agnostic, meaning that it can work with a department’s existing video surveillance infrastructure and scale to fit any department size and budget. With a video analytics tool like BriefCam, small and mid-sized police departments like Hartford PD can accelerate investigations, apprehend criminals, protect the public and gain a unique view of their community just as effectively as big city departments with more substantial budgets and much larger officer corps.
As a former police officer, O’Hare understands the skepticism with which police agencies eye new tools. “There's much noise in the analytics market today and historically video analytics solutions of the past didn't work in the way that the end user had anticipated,” said O’Hare. “But if you ever want to become a world-class, smart city, you're going to need to know what's happening around you, from traffic flow to crime. BriefCam can help you do that.”