A hypergraph-cut based denoising algorithm for ICCD image

ICCD image is captured in extremely low-light conditions by ICCD camera. It often contains spatially clustered noise, thus, general filtering methods do not work well. We present a denoising algorithm to remove the noise in the ICCD image. First, we over-segment the ICCD image into flat patches. Second, we classify the patches into noisy patches and noise-free patches based on hypergraph cut method. Then the noise-free patches are recovered by the general BM3D, since they do not contain clustered noise. The noisy patches are recovered by diminishing identified clustered noise from the noisy patches. After that, we could get the recovered ICCD image. Experiments are conducted on a set of ICCD images, which shows that the proposed algorithm remove the randomly clustered noise and preserve the textural information in the ICCD images.