Wide-band image guided visible-band image enhancement

In this paper, we present a new algorithm which enhances an image quality of the visible-band image in low light conditions by using the wide-band (400~1000 nm) image. In our proposed algorithm, the high signal-to-noise ratio (SNR) capability of the wide-band image is utilized, and the bivariate shrinkage and the multiresolution bilateral filter are incorporated with our joint filtering scheme. Our proposed method employs the cross channel paradigm to further reduce chromatic low frequency noise. Detail transfer is performed by using a mapping function on the wavelet domain. The experimental result with real image will be given in order to illustrate the effectiveness of the proposed algorithm.