A posteriori restoration of block transform-compressed data

Summary form only given; substantially as follows. The NASA/JPL Galileo spacecraft uses lossy data compression for the transmission of its science imagery over the low-bandwidth communication system. The technique for image compression is a block transform technique based on the integer cosine transform (ICT), a derivative of the JPEG image compression standard where a computationally efficient integer cosine transform implementation replaces JPEG's discrete cosine transform (DCT). JPEG and ICT are examples of block transform-based compression schemes. The compression is achieved by quantizing the transformed data. The distortion characteristics of block transform-based compression techniques are understandable in terms of the properties of the transform basis functions and the transform coefficient quantization error [Boden, 1995]. For scientific applications such as Galileo, it is particularly desirable to mitigate the quantization distortion in the decompressed image to enhance science return. Galileo's limited computational resources preclude additional processing, hence attempts at data restoration are limited to a posteriori processing. The present authors consider two known a posteriori enhancement techniques, transform coefficient adjustment [Ahumada Jr. and Horng], and low-pass filtering [Reeve III and Lim]. These techniques are adapted to ICT restoration based on a quantitative model of distortion statistics, and combined to achieve significant objective and subjective improvement in restored image fidelity compared with original data.