Image and video analysis with crowd’s consensus is one of the strong use cases of the Eventum platform. That is why on 28th of February (19:00 GMT+1) the 8th live demo event will start on Eventum — you will have to analyse video (a security camera feed) and report when any suspicious activity happens.
The problem is, that when something happens — that needs to be then reviewed on the tapes (e.g. break-in, robbery, vandalism …), hours, days or even weeks of video feeds need to be manually reviewed to find that specific event. Because every small movement (e.g. cat running across the street, trees moving in the wind, …) turns on the motion sensor, even the sophisticated systems do not make this an easy job.
To help with the manual review process, many companies and individuals use different built-in or 3rd-party AI systems to find the suspicious events and to filter out non-suspicious activity from the feed. Most of this AI systems, unfortunately, fail, because security camera video feeds are usually of low quality and have low framerate, while events itself are hard to distinguish for even the most state-of-the-art AI algorithms.
But, the human mind and its semantic segmentation capability are simply incredible. It is very easy for us to understand and recognize what is going on in the video. Of course, a single human can be wrong, and this represents a single point of failure — which is luckily solved by the Wisdom of Crowds or a SchellingPoint principle implemented in Eventum (learn more about why and how this works in the Whitepaper).
Eventum is a perfect solution to the above problem, where crowd detects and analyses the video, then if the consensus is made, the result is sent as a real-time API to any security company or individual in the world.
Of course, the security camera video feed is only a single use case, and there are many more in the image and video analysis space — see website, lightpaper and our past live demo events.
