Discrete Atomic Compression with Different Structures of Discrete Atomic Transform: Efficiency Comparison and Perspectives of Application to Digital Images Privacy Protection

Abstract

Discrete atomic compression (DAC) is an algorithm for lossy compression of digital images. Discrete atomic transform (DAT) is a core of this algorithm. An important feature of DAT is the ability to vary its structure, as well as the atomic function on which it is based. In this paper, we consider DAC with several different structures of the DAT-procedure. We show that such changes have no significant effect on efficiency of the algorithm in terms of compression ratio and loss of quality measured by maximum absolute deviation and root mean square error. This feature opens up wide opportunities for application of DAC to digital image privacy protection. We also compare DAC with JPEG. It is shown that DAC with different structures of DAT provides higher memory savings with the same quality of the result.