Degradation-invariant Enhancement of Fundus Images via Pyramid Constraint Network

TL;DR

This paper introduces PCE-Net, a pyramid constraint network that enhances fundus images by learning degradation-invariant features, reducing the need for clinical data and improving robustness in clinical scenarios.

Contribution

The paper proposes a novel pyramid constraint method and a degradation-invariant network for fundus image enhancement, addressing data scarcity and robustness issues.

Findings

PCE-Net outperforms state-of-the-art methods in enhancement quality.

The method improves segmentation accuracy on degraded fundus images.

Ablation studies validate the effectiveness of the pyramid constraint.

Abstract

As an economical and efficient fundus imaging modality, retinal fundus images have been widely adopted in clinical fundus examination. Unfortunately, fundus images often suffer from quality degradation caused by imaging interferences, leading to misdiagnosis. Despite impressive enhancement performances that state-of-the-art methods have achieved, challenges remain in clinical scenarios. For boosting the clinical deployment of fundus image enhancement, this paper proposes the pyramid constraint to develop a degradation-invariant enhancement network (PCE-Net), which mitigates the demand for clinical data and stably enhances unknown data. Firstly, high-quality images are randomly degraded to form sequences of low-quality ones sharing the same content (SeqLCs). Then individual low-quality images are decomposed to Laplacian pyramid features (LPF) as the multi-level input for the enhancement. Subsequently, a feature pyramid constraint (FPC) for the sequence is introduced to enforce the PCE-Net to learn a degradation-invariant model. Extensive experiments have been conducted under the evaluation metrics of enhancement and segmentation. The effectiveness of the PCE-Net was demonstrated in comparison with state-of-the-art methods and the ablation study. The source code of this study is publicly available at https://github.com/HeverLaw/PCENet-Image-Enhancement.