Section: Application Domains

Video Analytics

Stars has long-lasting experience in video analytics for security, an area of paramount importance due to the high need for public security, accompanied by the availability of a vast amount of cameras. Related applications range from public safety, smart retail, crowd management to maintenance of secure sites. One current challenge of video analytics is extracting the meaningful information from big video data in an efficient manner, with the goal of high-level semantic descriptions of the activities occurring in the area under surveillance.

Video analytics covers research directions including tracking, object detection, object classification, behavior recognition, pose estimation, and semantic scene modeling. Deep neural networks have brought to the fore great progress in the area of large-scale video analytics and specifically in object recognition and action recognition. However, many challenges remain open, mainly in the context of real-life surveillance scenarios and specifically concerning the robustness, adaptability, and scalability. Prominent open challenges comprise people detection and tracking in crowded environments, as well as detecting abnormal activities, which is often associated with processing of very large video data in the presence of noise and occlusions.

Stars has accumulated a well selected network of industrial partners ranging from end-users, integrators and software editors to provide data, objectives, evaluation and funding.