PROJECT TITLE :

Bio-inspired Image Enhancement derived from a 'rank order coding’ model

ABSTRACT:

In this study, the authors propose a brand new methodology to reinforce image data, based on wavelet decomposition and original partial reconstruction of image. This reconstruction called 'asynchronous reconstruction' isn't carried out in the identical manner as the standard sequential one. It's based mostly on rank order coding. Of course, while sequential reconstruction is to add all or a part of the responses obtained for each scale of 'coarse to fine' decomposition, asynchronous reconstruction tries to be nearer to human brain that uses a limited variety of frequency channels. Actually, when wavelet decomposition, responses are sorted from high down for every pixel of the image. Final asynchronous reconstruction for each pixel is obtained by adding a chosen range of wavelet responses, beginning by the most response. Thus, at a given level of reconstruction, the pixel values don't come from the identical frequency channels. The interest of this method has been tested on a face verification task using the IV2 biometric database. Stopping criterion for reconstruction will be a relentless number of wavelet responses to use, but an adaptive method has been also investigated. 3 criteria are explored: commonplace deviation, entropy and lost edges ratio.