RSAPower: Random Style Augmentation Driven Structure Perception Network for Generalized Retinal OCT Fluid Segmentation.

in IEEE transactions on medical imaging by Chenggang Lu, Zhitao Guo, Dan Zhang, Lei Mou, Jinli Yuan, Shaodong Ma, Da Chen, Yitian Zhao, Kewen Xia, Jiong Zhang

TLDR

  • RSAPower is a novel method for enhancing the generalization ability of fluid perception networks for retinal fluid segmentation in OCT images.
  • It achieves superior performance compared to state-of-the-art methods and demonstrates strong generalization ability.

Abstract

Optical Coherence Tomography (OCT) imaging is extensively utilized for non-invasive observation of pathological conditions, such as retinal fluid-associated diseases. Accurate fluid segmentation in OCT images is therefore critical for quantifying disease severity and aiding clinical decision-making. However, achieving precise segmentation remains challenging due to pathological variations in shape and size, uncertain boundaries, and low contrast of fluid. Most importantly, variability in OCT image styles across different vendors and centers significantly affects fluid segmentation, leading to poor generalization to unseen domains. To address this, we propose a novel method, RSAPower, to enhance the generalization ability of fluid perception networks via style augmentation for retinal fluid segmentation. Specifically, RSAPower comprises a plug-and-play random style transform augmentation (RSTAug) module and a novel fluid perception network (FLPNet) for end-to-end training. The RSTAug module generates new random-style data from the source domain, preserving realistic pathological and structural features. The FLPNet benefits from a novel hybrid structure attention (HSA) module to perceive fluid's spatial features and long-range dependence. Furthermore, FLPNet adapts to the diverse augmented data through a saliency-guided multi-scale attention (SGMA) block, boosting its segmentation performance. We validate RSAPower against various state-of-the-art methods using two publicly available datasets, Retouch and Kermany. Experimental results demonstrate the proposed method's superior generalization ability and effectiveness in fluid segmentation.

Overview

  • This study proposes a novel method, RSAPower, to enhance the generalization ability of fluid perception networks for retinal fluid segmentation.
  • RSAPower comprises a plug-and-play random style transform augmentation (RSTAug) module and a novel fluid perception network (FLPNet) for end-to-end training.
  • The method aims to address the challenge of accurate fluid segmentation in Optical Coherence Tomography (OCT) images due to pathological variations, uncertain boundaries, and low contrast.

Comparative Analysis & Findings

  • The proposed method, RSAPower, outperforms various state-of-the-art methods in fluid segmentation using two publicly available datasets, Retouch and Kermany.
  • RSAPower achieves superior generalization ability due to its ability to adapt to diverse augmented data through a saliency-guided multi-scale attention (SGMA) block.
  • The novel hybrid structure attention (HSA) module in FLPNet enables the perception of fluid's spatial features and long-range dependence.

Implications and Future Directions

  • The proposed method has the potential to improve clinical decision-making and quantification of disease severity in retinal fluid-associated diseases.
  • Future research directions may focus on exploring novel augmentation techniques to further enhance generalization ability and adapting RSAPower to other medical imaging modalities.
  • Additional studies can investigate the application of RSAPower in real-world settings and evaluate its performance on larger and more diverse datasets.
Read Full Article