Efficient image fusion using visual salient features and cross-contrast with edge weakening guided image filter

摘要

A novel algorithm determined by Nonsubsampled Contourlet Transform (NSCT) with edge weakening guided filter (EWGIF) is proposed during this paper by choosing and combining feature of the initial pictures. First of all initial pictures are break down into band pass and low subband coefficients by applying the nonsubsampled contourlet transform. Visible salience features is that the 1st picture that a viewer sees once staring at a picture, as an example, local energy, contrast and gradient are visual salience features. Lowpass subband coefficients are obtained by employing a visual salience feature whereas Band pass subband coefficients are acquired by using cross contrast. A cross-contrast method is got by calculating a ratio of the local mean of band-pass coefficients as well as local mean of lowpass coefficients which are solved by visual salience feature. Comparison tests are evaluated on various picture sets, and results outcome exhibit that the proposed technique performs better in both subjective and objective qualities.