CycleGAN with a Blur Kernel for Deconvolution Microscopy: Optimal Transport Geometry
This work addresses the need for unsupervised deconvolution in microscopy, offering a more efficient alternative to existing CNN methods that rely on supervised training, though it is incremental in its adaptation of cycleGAN.
The paper tackles the problem of deconvolution microscopy without requiring matched high-resolution images for training by proposing an unsupervised cycleGAN with a linear blur kernel. The method achieves improved robustness and efficiency in network training, as confirmed by experimental results on simulated and real data.
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.