Retrieval Augmented Generation and Patch-Level Explainability applied to Medical Image Captioning with Small Visual Language Models

CaptionFuse: Training-Free Composed Product Retrieval Via VLM Intent Captioning and Weighted Embedding Fusion

Seeing, Describing, and Explaining Medical Image Understanding via Retrieval and Multimodal Learning

Textual Preprocessing and Multimodal Conditioning for Medical Image Captioning

Attention-supervised transformer for grounded image captioning

MedFG-VQA: Low-Frequency Memory and Graph Attention for Lightweight Medical VQA

A lightweight framework that leverages a memory bank to augment DCT-based low-frequency features and employs graph-enhanced cross-attention for effective visual-textual alignment is proposed, highlighting its efficiency and potential for clinical deployment.

Model-Intrinsic Attention as Explanation for Radiology Image Captioning