IIHT: Medical Report Generation with Image-to-Indicator Hierarchical Transformer

TL;DR

This paper introduces IIHT, a hierarchical transformer framework that improves medical report generation by combining image features with disease indicators, enhancing clinical accuracy and report fluency.

Contribution

The paper proposes a novel image-to-indicator hierarchical transformer that integrates disease indicators into report generation, addressing data imbalance and sequence correlation issues in medical data.

Findings

Outperforms state-of-the-art methods on multiple metrics

Enables real-world indicator modification by radiologists

Achieves fluent and clinically accurate report generation

Abstract

Automated medical report generation has become increasingly important in medical analysis. It can produce computer-aided diagnosis descriptions and thus significantly alleviate the doctors' work. Inspired by the huge success of neural machine translation and image captioning, various deep learning methods have been proposed for medical report generation. However, due to the inherent properties of medical data, including data imbalance and the length and correlation between report sequences, the generated reports by existing methods may exhibit linguistic fluency but lack adequate clinical accuracy. In this work, we propose an image-to-indicator hierarchical transformer (IIHT) framework for medical report generation. It consists of three modules, i.e., a classifier module, an indicator expansion module and a generator module. The classifier module first extracts image features from the input medical images and produces disease-related indicators with their corresponding states. The disease-related indicators are subsequently utilised as input for the indicator expansion module, incorporating the "data-text-data" strategy. The transformer-based generator then leverages these extracted features along with image features as auxiliary information to generate final reports. Furthermore, the proposed IIHT method is feasible for radiologists to modify disease indicators in real-world scenarios and integrate the operations into the indicator expansion module for fluent and accurate medical report generation. Extensive experiments and comparisons with state-of-the-art methods under various evaluation metrics demonstrate the great performance of the proposed method.