Skip to Main Content
Many bandwidth compression techniques which have been applied to imagery can be characterized as low pass filters. Higher compression rates yield images with reduced resolution or sharpness. In general, the sharpness in an image is a function of high contrast edges. Several studies have shown that, although the presence of these edges is important to the overall subjective quality of the image, their fidelity is not. High contrast edge information can be isolated in the upper bit plane (the most significant bit) of most types of imagery. Simple run length encoding of this bit plane can be used to preserve the location and approximate peak amplitude of the edge information at an overhead cost of less than 0.1 bit/pixel. When upper bit plane run length encoding is combined with standard transform or DPCM coding: the resultant hybrid technique provides images with subjective quality improvements of better than two to one. This hybrid approach has been demonstrated on several types of imagery. Current activity is centered on an automatic intensity remapping function which guarantees the upper bit plane contains the optimal amount of information to ensure maximum run length encoding efficiency.