Receiver architectures for real-time image-compression algorithms based on a feature-visibility model

Receiver architectures for a computationally simple image-compression scheme based on an explicit feature-visibility criterion are discussed. The compression algorithms use the feature-visibility criterion to remove image detail not visible under normal viewing conditions. The algorithms are of low complexity, use integer arithmetic operations, and are well suited to simple real-time implementation in hardware. The algorithms break the image into square regions of uniform intensity. A lossless code has been developed that efficiently encodes the region size and organization information. The receiver stages used to decode the lossless code, to reconstruct the compressed image, and to postfilter the image for smoothing are described.