No-inference image sharpness assessment based on wavelet transform and image saliency map
Hengjun Zhao · 2016
This paper combines the wavelet transform and image saliency map to construct an image sharpness index without reference image. The main idea of the proposed index is that the saliency areas in an image always get more attentions than non-saliency regions when human perceives the sharpness of the image. Specifically, an input image is firstly decomposed into three directional sub-bands by a separable discrete wavelet transform. Next, the saliency map of the inputted image is constructed by Itti's salience algorithm. Then, the log-energies of wavelet coefficients belonging to the saliency area of the input image are computed. The final scalar index corresponding to the image's overall sharpness is computed via a weighted average of these log-energies. The experiments show that, despite its simplicity, the proposed sharpness index is competitive with the current best-performance techniques for no-reference image sharpness estimation.