LWGNet: Learned Wirtinger Gradients for Fourier Ptychographic Phase Retrieval

TL;DR

LWGNet is a hybrid neural network that unrolls Wirtinger flow optimization for Fourier ptychographic phase retrieval, improving image quality especially with low-cost sensors and reducing computational complexity.

Contribution

It introduces LWGNet, a novel neural network unrolling of Wirtinger flow, combining physics-based modeling with deep learning for enhanced FPM reconstruction.

Findings

LWGNet outperforms traditional and deep learning methods in low-bit sensor scenarios.

Fewer unrolling stages achieve comparable or better results.

Improved performance demonstrated on real data.

Abstract

Fourier Ptychographic Microscopy (FPM) is an imaging procedure that overcomes the traditional limit on Space-Bandwidth Product (SBP) of conventional microscopes through computational means. It utilizes multiple images captured using a low numerical aperture (NA) objective and enables high-resolution phase imaging through frequency domain stitching. Existing FPM reconstruction methods can be broadly categorized into two approaches: iterative optimization based methods, which are based on the physics of the forward imaging model, and data-driven methods which commonly employ a feed-forward deep learning framework. We propose a hybrid model-driven residual network that combines the knowledge of the forward imaging system with a deep data-driven network. Our proposed architecture, LWGNet, unrolls traditional Wirtinger flow optimization algorithm into a novel neural network design that enhances the gradient images through complex convolutional blocks. Unlike other conventional unrolling techniques, LWGNet uses fewer stages while performing at par or even better than existing traditional and deep learning techniques, particularly, for low-cost and low dynamic range CMOS sensors. This improvement in performance for low-bit depth and low-cost sensors has the potential to bring down the cost of FPM imaging setup significantly. Finally, we show consistently improved performance on our collected real data.