CycleGAN with a Blur Kernel for Deconvolution Microscopy: Optimal Transport Geometry

TL;DR

This paper introduces an unsupervised cycleGAN model with a linear blur kernel for deconvolution microscopy, improving robustness and efficiency by leveraging optimal transport geometry, and demonstrating effectiveness on simulated and real data.

Contribution

It proposes a novel cycleGAN architecture with a single generator and a linear blur kernel, formulated as an optimal transport problem, for blind and non-blind deconvolution microscopy.

Findings

Effective on simulated data

Validated on real experimental data

Improves robustness and training efficiency

Abstract

Deconvolution microscopy has been extensively used to improve the resolution of the wide-field fluorescent microscopy, but the performance of classical approaches critically depends on the accuracy of a model and optimization algorithms. Recently, the convolutional neural network (CNN) approaches have been studied as a fast and high performance alternative. Unfortunately, the CNN approaches usually require matched high resolution images for supervised training. In this paper, we present a novel unsupervised cycle-consistent generative adversarial network (cycleGAN) with a linear blur kernel, which can be used for both blind- and non-blind image deconvolution. In contrast to the conventional cycleGAN approaches that require two deep generators, the proposed cycleGAN approach needs only a single deep generator and a linear blur kernel, which significantly improves the robustness and efficiency of network training. We show that the proposed architecture is indeed a dual formulation of an optimal transport problem that uses a special form of the penalized least squares cost as a transport cost. Experimental results using simulated and real experimental data confirm the efficacy of the algorithm.