Incomplete Modality Disentangled Representation for Ophthalmic Disease Grading and Diagnosis

TL;DR

This paper introduces a novel method called IMDR that disentangles features into modal-common and modal-specific components to improve ophthalmic disease diagnosis with incomplete multimodal data, addressing limitations of implicit representations and modality heterogeneity.

Contribution

The paper proposes the IMDR strategy that explicitly disentangles features and a joint proxy learning module, enhancing robustness and missing data reconstruction in multimodal ophthalmic diagnosis.

Findings

IMDR outperforms state-of-the-art methods on four datasets

Disentangled features improve missing modality reconstruction

Joint proxy learning reduces intra-modality redundancy

Abstract

Ophthalmologists typically require multimodal data sources to improve diagnostic accuracy in clinical decisions. However, due to medical device shortages, low-quality data and data privacy concerns, missing data modalities are common in real-world scenarios. Existing deep learning methods tend to address it by learning an implicit latent subspace representation for different modality combinations. We identify two significant limitations of these methods: (1) implicit representation constraints that hinder the model's ability to capture modality-specific information and (2) modality heterogeneity, causing distribution gaps and redundancy in feature representations. To address these, we propose an Incomplete Modality Disentangled Representation (IMDR) strategy, which disentangles features into explicit independent modal-common and modal-specific features by guidance of mutual information, distilling informative knowledge and enabling it to reconstruct valuable missing semantics and produce robust multimodal representations. Furthermore, we introduce a joint proxy learning module that assists IMDR in eliminating intra-modality redundancy by exploiting the extracted proxies from each class. Experiments on four ophthalmology multimodal datasets demonstrate that the proposed IMDR outperforms the state-of-the-art methods significantly.