CauDR: A Causality-inspired Domain Generalization Framework for Fundus-based Diabetic Retinopathy Grading
This addresses the domain shift issue in medical imaging for ophthalmologists, enabling more reliable computer-aided diagnosis, though it is incremental as it builds on causality methods for a specific application.
The paper tackled the problem of limited generalization in deep learning-based diabetic retinopathy grading across different imaging domains by proposing CauDR, a causality-inspired framework that eliminates spurious correlations, achieving state-of-the-art performance on a reorganized 4DR benchmark.
Diabetic retinopathy (DR) is the most common diabetic complication, which usually leads to retinal damage, vision loss, and even blindness. A computer-aided DR grading system has a significant impact on helping ophthalmologists with rapid screening and diagnosis. Recent advances in fundus photography have precipitated the development of novel retinal imaging cameras and their subsequent implementation in clinical practice. However, most deep learning-based algorithms for DR grading demonstrate limited generalization across domains. This inferior performance stems from variance in imaging protocols and devices inducing domain shifts. We posit that declining model performance between domains arises from learning spurious correlations in the data. Incorporating do-operations from causality analysis into model architectures may mitigate this issue and improve generalizability. Specifically, a novel universal structural causal model (SCM) was proposed to analyze spurious correlations in fundus imaging. Building on this, a causality-inspired diabetic retinopathy grading framework named CauDR was developed to eliminate spurious correlations and achieve more generalizable DR diagnostics. Furthermore, existing datasets were reorganized into 4DR benchmark for DG scenario. Results demonstrate the effectiveness and the state-of-the-art (SOTA) performance of CauDR.