A New Poisson Noisy Image Denoising Method Based on the Anscombe Transformation

In this paper, we propose a new denoising method for Poisson noise corrupted images based on the Anscombe variance stabilizing transformation (VST) with a new inversion. The VST is used to approximately convert a Poisson noise image into a Gaussian distributed image, so that the denoising methods aiming at Gaussian noise can be applied subsequently. The motivation for the new inversion originates from a main drawback existing in the Anscombe transformation: its efficiency degrades significantly when the pixel intensities of the observed images are very low due to the biased errors generated by its inverse transformation. Thus, we introduce a polynomial regression model in the sense of weighted least squares as an improvement for the inverse Anscombe transformation. Moreover, we incorporate our developed wavelet thresholding strategy for Gaussian noise into the proposed method. It is shown in the experimental analysis that this method is very competitive for Poisson