We continue the presentation of Contractors which were qualified to this stage of works. One of them is PreventUS A.I. The application was submitted by Ayvos A.S., a company on high-tech computer vision software industry that provides end-to-end solutions for several sectors and scenarios. What technology did they propose?
Object Counter is a method used to count the people detected in an area monitored by a camera. The PreventUs A.I. solution will be able to supply feature predictions that are used to calculate similarity scores between detection couplets. Those scores are then used to associate detections that belong to the same target.
IDs are assigned to similar detections, and different IDs are applied to detections that are not part of pairs. With AYVOS AODT (Any Object Detector and Tracker), any object in our real world will be able to be added for being detected and tracked in addition to known classifications like human, pedestrian, vehicle (*car, train, bus"), etc.
Videos from actual fields will be automatically collected via frame-cutting and cleaning infrastructure that is autonomous, and new item kinds may always be trained iteratively. The people who detected and followed up with AODT are counted in the regions monitored by cameras.
At least four of the six Consortia competing in the Solution Design phase will be qualified to the next stage, during which the Contractors will develop a first prototype of a technology. To learn more about the PCP approach, please visit this subpage.