About Opportunity:

The technology works by “prioritizing” certain components of a video scene, in particular human faces and foreground objects. These components are sent with higher quality, resulting in improved overall quality. This is of particular benefit to “face to face” applications, including the specialist application of sign language communication for deaf users.

We have successfully developed a prototype software system that demonstrates the Perceptually Optimised Video Compression concept that performs the following functions:

• Captures video from a camera (e.g. a webcam)
• Identifies the location of human faces in the video scene (face detection)
• Tracks the dominant human face in the scene (face tracking)
• Pre-filters (prioritizes) the scene in order to emphasise the face and de-emphasise other regions (pre-filtering). We have implemented two alternative pre-filters, incl a novel variable bilateral filter.
• Compressed (encodes) the video stream using RGU’s H.264 video encoder.
• Transports the compressed video stream via the internet
• Decompresses (decodes) the video stream.
• Displays the output on the PC screen.