I am interested in knowing more about denoising in 2-photon fluorescent microscopy acquisitions. In my lab usually, we combine 2 indicators (one in red channel and the second one in green as GCaMP6f). We extract small volumes of 12x12x5um (64x64x5 voxels) at high speed (10Hz). Most preprocessing pipelines assume that the noise is Gaussian (like with the estimator of noise variation on skimage) but I wonder if the noise could be decomposed in more concrete components. Is there any other type of noise (like Poisson noise) relevant for 2 photon microscopy? For instance, the microscope sometimes adds some structural artifacts to the acquisition. Is there any good review about the kinds of noise that I can follow? What are the best metrics to check if an algorithm is denoising correctly?
Thanks in advance