GCS-M3VLT: Guided Context Self-Attention based Multi-modal Medical Vision Language Transformer for Retinal Image Captioning

TL;DR

This paper introduces GCS-M3VLT, a novel vision-language transformer that effectively combines visual and textual features for retinal image captioning, especially in data-scarce scenarios, improving caption quality.

Contribution

It proposes a guided context self-attention mechanism that enhances multi-modal feature integration in retinal image captioning models with limited labeled data.

Findings

Achieved a 0.023 BLEU@4 improvement on DeepEyeNet dataset.

Demonstrated qualitative improvements in medical caption generation.

Effective in data-scarce medical imaging scenarios.

Abstract

Retinal image analysis is crucial for diagnosing and treating eye diseases, yet generating accurate medical reports from images remains challenging due to variability in image quality and pathology, especially with limited labeled data. Previous Transformer-based models struggled to integrate visual and textual information under limited supervision. In response, we propose a novel vision-language model for retinal image captioning that combines visual and textual features through a guided context self-attention mechanism. This approach captures both intricate details and the global clinical context, even in data-scarce scenarios. Extensive experiments on the DeepEyeNet dataset demonstrate a 0.023 BLEU@4 improvement, along with significant qualitative advancements, highlighting the effectiveness of our model in generating comprehensive medical captions.