Cost-Sensitive Distribution Alignment for Improving Automated Medical Diagnosis Using Retinal Fundus Photography

Domain adaptation (DA) has emerged as a promising approach to address the domain shift problem in deep learning for automated medical diagnosis. However, current approaches often overlook the imbalanced nature of different categories and primarily focus on aligning the distributions of the source and target domains globally. This oversight leads to suboptimal performance on imbalanced target datasets, as the alignment process becomes dominated by the majority class during adaptation. To tackle this limitation, this paper proposes a novel domain adaptation method called cost-sensitive distribution alignment (CSDA), which aims to enhance deep learning performance on target data. The proposed approach involves collecting a limited-sized dataset from the target domain and utilising CSDA to bridge the domain gap between the source and target domains. Specifically, to address class imbalance, cost-sensitive learning is incorporated into the distribution alignment process, giving more emphasis to the misalignment cost associated with minority class samples. More importantly, CSDA aligns the cross-domain projection distribution in the feature space with the ideal geometric distribution derived from the ground-truth labels. Unlike existing methods that directly minimise or employ adversarial learning to reduce the distribution discrepancy between the source and target domains, our proposed CSDA method focuses on minimising the semantic relationship misalignment among cross-domain samples. Experimental results on the detection of high myopia and myopic macular degeneration (MMD) demonstrate the superiority of CSDA over state-of-the-art methods. The results provide empirical evidence of CSDA’s efficacy in enhancing automated medical diagnosis.

I. Introduction

Vision, as the primary means of perceiving the world, is essential in all facets of life. Regrettably, eye conditions and vision impairment are prevalent and pose a significant challenge to eye care, particularly in low- and middle-income nations. The World Report on Vision by the World Health Organisation (WHO) indicates that globally, at least 2.2 billion individuals suffer from vision impairment, and of these, at least 1 billion cases could have been prevented or remain unaddressed [1]. Early diagnosis and treatment are crucial in preventing widespread vision conditions and impairments. Fundus retinal photography, a non-invasive imaging method for capturing colour images of the interior surface of the eye, is widely used for detecting eye disorders and monitoring their progress over time, as it is quick to complete and does not require any invasive procedures. However, the interpretation of fundus photographs is dependent on the expertise of experienced specialists to identify disease pathology. To increase speed and scale in interpretation, artificial intelligence (AI), particularly deep learning [2]–[4], has been applied to detect major ophthalmic diseases from high-quality retinal fundus images.