A Python Flask web application using AI (deep learning) for object detection and recognition in video streams, for their further usage in areas such as advertisements.
In this web application, the user has the following possibilities: one can upload a recording of a sport event (match, meeting, interview etc.) to our server. Subsequently, our pre-trained Neural Network detects and recognizes logos (implemented at this moment), as well as faces and texts (planned in the future), which then are linked to respective advertisement information, and the final results are presented to the clients. The video can be paused and rewound to an arbitrary position using the slider on the player page. The user can disable the detection of logos or texts at any time, to facilitate faster processing. Editing of logo-advertisement bindings, as well as manual correction and training of the neural network are possible through an admin panel, currently hosted at admin.cutrack.net