Optimized and Privacy-aware Multi-view Streaming

The project will deliver optimized solutions enabling our vision of a new immersive streaming service that (1) uses camera recordings from relatively inexpensive multi-camera settings to provide a service that allows users to (2) view a scene/event from any angle, (3) zoom in on specific features during a movie, concert, or game, (4) request facial features of detected faces to be manipulated to the user´s preference (e.g., by changing their age, gender, or make them happier/sadder), and the system at all times still ensures that (5) the identity of the actual people visible in the final video stream viewed by the user should be protected and yet realistic.

To achieve our objectives, we will design solutions that push the state-of-the-art in three research areas: (1) multi-view and free viewpoint streaming using simultaneous video footage from several relatively inexpensive cameras, (2) next generation privacy enhancing technologies, and (3) optimized end-to-end service delivery that consider where and how the processing is done.

For each area, we will produce both fundamental results and practical solutions. Our solutions will aim to preserve the naturalness of faces and optimize the user´s quality of experience (QoE).

Enabling and providing optimized delivery of these services over the Internet in close-to realtime is a challenging and important problem that has the potential to disrupt both the entertainment industry and how we do business and interact with people.