6533b871fe1ef96bd12d0da9

RESEARCH PRODUCT

Comparative study of multi-2D, Full 3D and hybrid strategies for multi/hyperspectral image compression

Alamin MansouriJonathan DelcourtYvon VoisinTadeusz Sliwa

subject

Image codingTexture compressionbusiness.industryCompression (functional analysis)Hyperspectral image compressionPrincipal component analysisPattern recognitionArtificial intelligencebusinessDecorrelationMathematicsData compression

description

In this paper, we investigate appropriate strategies for multi/hyperspectral image compression. In particular, we compare the classic multi-2D compression strategy and two different implementations of 3D strategies (Full 3D and hybrid). All strategies are combined with a PCA decorrelation stage to optimize performance. For multi-2D and hybrid strategies, we propose a weighted version of PCA. Finally, for consistent evaluation, we propose a larger comparison framework than the conventionally used PSNR. The results are significant and show the weaknesses and strengths of each strategy.

yearjournalcountryeditionlanguage
2009-09-012009 Proceedings of 6th International Symposium on Image and Signal Processing and Analysis
https://doi.org/10.1109/ispa.2009.5297748