Lossless image compression using autosophy networks

Parallel autosophy networks may soon provide an entirely new technique for transmitting HDTV or for storing graphic images in a computer. The new compression technique will not distort the images or produce visual artifacts. The transmission bandwidth in television or the storage requirements in computer disc files is entirely independent of the image size, image resolutions or frame rates. It only depends on “novelty” or “movement” within the images or how similar the images are to each other. In teleconferencing or graphics animation only a small area in the image changes between frames. These changing portions in the images are selectively encoded for transmission or for storage in a peculiar mathematical omni-dimensional hyperspace library in which no image portion is stored twice. The more images are already stored the less memory space is required to store additional images leading to rapid memory saturation and enormous data compression. For television the input images are compared like a jigsaw puzzle with all the images in the hyperspace library. The largest matching image portions are encodes into “superpixel” codes which may represent any large portion of the image. For computer graphics storage the input images are absorbed in the hyperspace library and represented by single output codes per frame to the computer. A software simulation in our laboratory, using C++ language in an IBM PC, has demonstrated the enormous potential of these new image compression techniques. Hardware systems will soon follow