Integrated Wavelet, Fractal and Neural Network Video Compression System

Current data rates for use on IR/EO LOROPS or ATARS platforms can exceed 120Mb/s, and, thus, a digital image sequence must be coded prior to transmission and storage. To remove the redundant information, which always exists in image sequences, most video compression techniques use 2D coding to achieve spatial compression and motion compensated difference coding in the time dimension. We propose to apply the wavelet transform, fractals, and neural networks independently in all three dimensions to reduce the computational complexity of coding while achieving high rates of compression. A 3D video coding method which combines simultaneous spatial and temporal domain predictions in an attempt to code effectively the 3D wavelet tranform coefficent of image sequence will be developed as part of our project. Our architecture will produce the best compression integrating three independent methods.