An adaptive window approach for Poisson noise reduction and structure preserving in confocal microscopy

In domains like confocal microscopy, the imaging process is based on detection of photons. It is established the additive Gaussian noise model is a poor description of the actual photon-limited image recording, compared with that of a Poisson process. This motivates the use of restoration methods optimized for Poisson noise distorted images. In this paper, we propose a novel restoration approach for Poisson noise reduction and discontinuities preservation in images. The method is based on a local modeling of the image, with an adaptive choice of a window around each pixel in which the applied model fits the data well. The restoration technique associates with each pixel the weighted sum of data points within the window. It is worth noting the proposed technique applied to confocal microscopy is data-driven and does not require the hand tuning of parameters.