Textual Inversion and Self-supervised Refinement for Radiology Report Generation

TL;DR

This paper introduces TISR, a novel method combining textual inversion and self-supervised refinement to improve radiology report generation by bridging modality gaps and enhancing report fidelity.

Contribution

The proposed TISR method effectively addresses modality gap and report content constraints, offering a plug-and-play solution that improves report quality.

Findings

Significant performance improvements on public datasets

Effective reduction of modality gap in report generation

Enhanced report fidelity through contrastive learning

Abstract

Existing mainstream approaches follow the encoder-decoder paradigm for generating radiology reports. They focus on improving the network structure of encoders and decoders, which leads to two shortcomings: overlooking the modality gap and ignoring report content constraints. In this paper, we proposed Textual Inversion and Self-supervised Refinement (TISR) to address the above two issues. Specifically, textual inversion can project text and image into the same space by representing images as pseudo words to eliminate the cross-modeling gap. Subsequently, self-supervised refinement refines these pseudo words through contrastive loss computation between images and texts, enhancing the fidelity of generated reports to images. Notably, TISR is orthogonal to most existing methods, plug-and-play. We conduct experiments on two widely-used public datasets and achieve significant improvements on various baselines, which demonstrates the effectiveness and generalization of TISR. The code will be available soon.