Enhancing Digital Hologram Reconstruction Using Reverse-Attention Loss for Untrained Physics-Driven Deep Learning Models with Uncertain Distance

TL;DR

This paper introduces a novel reverse-attention loss for untrained physics-based deep learning in digital holography, effectively addressing autofocusing challenges caused by uncertain object distances, and achieving near-optimal reconstruction accuracy.

Contribution

It proposes the first reverse-attention loss method that improves autofocus in untrained physics-driven deep learning models for hologram reconstruction.

Findings

Outperforms existing methods in efficiency and accuracy.

Achieves near-perfect reconstruction comparable to known object distances.

Demonstrates theoretical and experimental superiority over traditional approaches.

Abstract

Untrained Physics-based Deep Learning (DL) methods for digital holography have gained significant attention due to their benefits, such as not requiring an annotated training dataset, and providing interpretability since utilizing the governing laws of hologram formation. However, they are sensitive to the hard-to-obtain precise object distance from the imaging plane, posing the $\textit{Autofocusing}$ challenge. Conventional solutions involve reconstructing image stacks for different potential distances and applying focus metrics to select the best results, which apparently is computationally inefficient. In contrast, recently developed DL-based methods treat it as a supervised task, which again needs annotated data and lacks generalizability. To address this issue, we propose $\textit{reverse-attention loss}$, a weighted sum of losses for all possible candidates with learnable weights. This is a pioneering approach to addressing the Autofocusing challenge in untrained deep-learning methods. Both theoretical analysis and experiments demonstrate its superiority in efficiency and accuracy. Interestingly, our method presents a significant reconstruction performance over rival methods (i.e. alternating descent-like optimization, non-weighted loss integration, and random distance assignment) and even is almost equal to that achieved with a precisely known object distance. For example, the difference is less than 1dB in PSNR and 0.002 in SSIM for the target sample in our experiment.