Title :
Evolving matched forward and inverse transforms for improved lossy compression of images from mars
Author :
Moore, Frank ; Babb, Brendan ; Aldridge, Shawn ; Peterson, Michael R.
Author_Institution :
Math. Sci. Dept., Univ. of Alaska Anchorage, Anchorage, AK, USA
Abstract :
Space missions currently rely upon wavelet-based techniques for image compression and reconstruction. Recent research has established a methodology for using evolutionary computation to optimize image reconstruction transforms that outperform wavelets under lossy conditions. This paper describes new research that builds upon previous work by evolving matched filter pairs that outperform wavelets and previously evolved reconstruction transforms for lossy compression and reconstruction of images transmitted from deep space. The new transforms significantly reduce data loss due to quantization error for this class of images, while continuing to match the compression capabilities of wavelets.
Keywords :
Mars; astronomical image processing; evolutionary computation; image coding; image reconstruction; inverse transforms; matched filters; quantisation (signal); wavelet transforms; Mars; data loss reduction; deep space image transmission; evolutionary computation; image reconstruction transform; inverse transforms; lossy image compression; matched filter pairs; matched forward transform; quantization error; space mission; wavelet-based technique; Image coding; Image reconstruction; Multiresolution analysis; Quantization; Training; Wavelet transforms; genetic algorithms; image compression; quantization;
Conference_Titel :
Systems, Man, and Cybernetics (SMC), 2011 IEEE International Conference on
Conference_Location :
Anchorage, AK
Print_ISBN :
978-1-4577-0652-3
DOI :
10.1109/ICSMC.2011.6083712
