作者: Jonathan Delcourt , Alamin Mansouri , Tadeusz Sliwa , Yvon Voisin
关键词: Decorrelation 、 Set partitioning in hierarchical trees 、 Wavelet 、 Pattern recognition 、 Pixel 、 Artificial intelligence 、 Mathematics 、 Multispectral image 、 Hyperspectral imaging 、 Principal component analysis 、 Wavelet transform
摘要: In this paper, we investigate different approaches for multi/hyperspectral image compression. particular, compare the classic multi-2D compression approach and two implementations of 3D (full hybrid) with regards to variations in spatial spectral dimensions. All are combined a weighted Principal Component Analysis (PCA) decorrelation stage optimize performance. For consistent evaluation, propose larger comparison framework than conventionally used PSNR, including eight metrics divided into three families. The results show weaknesses strengths each approach.