Denoising of two-photon fluorescence images with block-matching 3D filtering.

Two-photon florescence imaging is widely used to perform morphological analysis of subcellular structures such as neuronal dendrites and spines, astrocytic processes etc. This method is also indispensable for functional analysis of cellular activity such as Ca2+ dynamics. Although spatial resolution of laser scanning two-photon system is greater than that of confocal or wide field microscope, it is still diffraction limited. In practice, the resolution of the system is more affected by its signal-to-noise ratio (SNR) than the diffraction limit. Thus, various approaches aiming to increase the SNR in two-photon imaging are desirable and can potentially save on building costly super-resolution imaging system. Here we analyze the statistics of noise in the two-photon florescence images of hippocampal astrocytes expressing genetically encoded Ca2+ sensor GCaMP2 and show that it can be reasonably well approximated using the same models which are used for describing noise in images acquired with digital cameras. This allows to use denoising methods available for wide field imaging on two-photon images. Particularly we demonstrate that the Block-Matching 3D (BM3D) filter can significantly improve the quality of two-photon fluorescence images so small details such as astrocytic processes can be easier identified. Moreover, denoising of the images with BM3D yields less noisy Ca2+ signals in astrocytes when denoising of the images with Gaussian filter.

BM3D frames and variational image deblurring.

A family of the block matching 3-D (BM3D) algorithms for various imaging problems has been recently proposed within the framework of nonlocal patchwise image modeling , . In this paper, we construct analysis and synthesis frames, formalizing BM3D image modeling, and use these frames to develop novel iterative deblurring algorithms. We consider two different formulations of the deblurring problem, i.e., one given by the minimization of the single-objective function and another based on the generalized Nash equilibrium (GNE) balance of two objective functions. The latter results in the algorithm where deblurring and denoising operations are decoupled. The convergence of the developed algorithms is proved. Simulation experiments show that the decoupled algorithm derived from the GNE formulation demonstrates the best numerical and visual results and shows superiority with respect to the state of the art in the field, confirming a valuable potential of BM3D-frames as an advanced image modeling tool.