Visual Attention Based Hyperspectral Imagery Visualization

In this paper, with physiologic structure and cognitive characteristic of the human vision system used, a model of ROI Visualization for Hyperspectral Images (HSI) which based on bottom-up visual attention is proposed. It has capability of visually partial selectivity in the human vision system. This method can directly process and display HSI, and take into account both spatial and spectral information. Meanwhile, it does not depend on the original data´s dimensional reduction method. This approach is applied to the real HSIs, and the exciting outputs show it is effective.