A new quantization method that uses the criterion of preserving sample absolute moments is presented. This is based on the same basic idea for block truncation coding of Delp and Mitchell but it is simpler in any practical implementation. Moreover, output equations are those for a two-level nonparametric minimum mean square error quantizer when the threshold is fixed to the sample mean. The application of this method to single frame color images is developed. A color image coding system that uses absolute moment block truncation coding of luminance and chroma information is presented. Resulting color images show reasonable performance with bit rates as low as 2.13 bits/pixel.

A technique for bandwidth compression coding of sequential digitized video imagery is presented. The method uses adaptive one-bit quantization over small blocks of space and time. The system includes global registration and motion detection for optimal hit assignments. A comparsion is made with a conditional replenishment technique using a motion compensated predictor described in the literature. The method presented produces reconstructed video sequences filmed from low altitude aircraft with slight degradation which is acceptable for many remote sensing applications at 0.9 bit/pixel data rate. Robust behavior in the presence of channel noise is demonstrated.