Denoising images with Poisson noise using M-transformation in wavelet domain
Yeqiu Li, Jianming Lu, Ling Wang, and Takashi Yahagi
Journal of Signal Processing, vol.11, no.4, pp.281-284, July, 2007. [pdf document]

<Abstract>

A Poisson process is a useful statistical model for a wide variety of counting problems in a number of different fields. In this paper, we study a new approach of removing Poisson noise from a degraded image. This method widens the BayesShrink algorithm into a double-density discrete wavelet transform (DWT) domain using M-transformation in preprocessing. By M-transformation, Poisson noise is converted into a small-amplitude random signal. Since the BayesShrink method is based on the assumption that the removed noise is Gaussian white noise, we apply it to the M-Transformed image to remove Poisson noise effeciently.