Data Compression of Time-varying Images

Abstract: 
Current methods of transmitting time-varying images involve encoding the digital image information into a set of still frames. To reduce the bandwidth necessary to transmit the full image, compression techniques are used to eliminate statistically redundant information in the pixel intensity distribution in each image frame and between each pair of successive image frames. These spatial compression techniques have now reached their compression limits with compression ratios of several tens.

Researchers have developed a method of time-varying image compression which provides compression ratios of several hundred. Unlike spatial compression techniques, information on the pixel intensity values are only transmitted for intermittent frames. Between the intermittent frames, the invention compresses the image motion rather than the spatial pixel intensity distribution. The data compression exploits the redundancy of information in a sequence of frames associated with the motions of the objects in the sequence. Combined with current technology of high-speed compression of still images, this invention achieves compression ratios of over 200 in cinematic sequences of digital frames.

Advantage:
The invention enables good quality color and black & white video communication, both for high definition TV and for teleconferencing over telephone lines. It also enables significant dose reduction in medical cinefluoroscopy, and great reductions in storage space for cinematic or video imagery. Many other applications are found in communications, imaging, and graphics.