Domain Generalization with Fourier Transform and Soft Thresholding

TL;DR

This paper introduces a novel Fourier-transform-based domain generalization method using soft thresholding to improve neural network performance on unseen domains, demonstrated on retinal fundus image segmentation.

Contribution

It proposes a new soft-thresholding technique in the Fourier domain to reduce background interference, enhancing domain generalization in medical image segmentation.

Findings

Outperforms conventional Fourier-based methods in segmentation accuracy.

Reduces background interference in Fourier amplitude spectrum.

Improves generalization to unseen domains in retinal image segmentation.

Abstract

Domain generalization aims to train models on multiple source domains so that they can generalize well to unseen target domains. Among many domain generalization methods, Fourier-transform-based domain generalization methods have gained popularity primarily because they exploit the power of Fourier transformation to capture essential patterns and regularities in the data, making the model more robust to domain shifts. The mainstream Fourier-transform-based domain generalization swaps the Fourier amplitude spectrum while preserving the phase spectrum between the source and the target images. However, it neglects background interference in the amplitude spectrum. To overcome this limitation, we introduce a soft-thresholding function in the Fourier domain. We apply this newly designed algorithm to retinal fundus image segmentation, which is important for diagnosing ocular diseases but the neural network's performance can degrade across different sources due to domain shifts. The proposed technique basically enhances fundus image augmentation by eliminating small values in the Fourier domain and providing better generalization. The innovative nature of the soft thresholding fused with Fourier-transform-based domain generalization improves neural network models' performance by reducing the target images' background interference significantly. Experiments on public data validate our approach's effectiveness over conventional and state-of-the-art methods with superior segmentation metrics.