Online Iterative Self-Alignment for Radiology Report Generation

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Xiao, Ting, Shi, Lei, Zhang, Yang, Yang, HaoFeng, Wang, Zhe, Bai, Chenjia
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912383633457152
author Xiao, Ting
Shi, Lei
Zhang, Yang
Yang, HaoFeng
Wang, Zhe
Bai, Chenjia
author_facet Xiao, Ting
Shi, Lei
Zhang, Yang
Yang, HaoFeng
Wang, Zhe
Bai, Chenjia
contents Radiology Report Generation (RRG) is an important research topic for relieving radiologist' heavy workload. Existing RRG models mainly rely on supervised fine-tuning (SFT) based on different model architectures using data pairs of radiological images and corresponding radiologist-annotated reports. Recent research has shifted focus to post-training improvements, aligning RRG model outputs with human preferences using reinforcement learning (RL). However, the limited data coverage of high-quality annotated data poses risks of overfitting and generalization. This paper proposes a novel Online Iterative Self-Alignment (OISA) method for RRG that consists of four stages: self-generation of diverse data, self-evaluation for multi-objective preference data,self-alignment for multi-objective optimization and self-iteration for further improvement. Our approach allows for generating varied reports tailored to specific clinical objectives, enhancing the overall performance of the RRG model iteratively. Unlike existing methods, our frame-work significantly increases data quality and optimizes performance through iterative multi-objective optimization. Experimental results demonstrate that our method surpasses previous approaches, achieving state-of-the-art performance across multiple evaluation metrics.
format Preprint
id arxiv_https___arxiv_org_abs_2505_11983
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Online Iterative Self-Alignment for Radiology Report Generation
Xiao, Ting
Shi, Lei
Zhang, Yang
Yang, HaoFeng
Wang, Zhe
Bai, Chenjia
Computer Vision and Pattern Recognition
Artificial Intelligence
Radiology Report Generation (RRG) is an important research topic for relieving radiologist' heavy workload. Existing RRG models mainly rely on supervised fine-tuning (SFT) based on different model architectures using data pairs of radiological images and corresponding radiologist-annotated reports. Recent research has shifted focus to post-training improvements, aligning RRG model outputs with human preferences using reinforcement learning (RL). However, the limited data coverage of high-quality annotated data poses risks of overfitting and generalization. This paper proposes a novel Online Iterative Self-Alignment (OISA) method for RRG that consists of four stages: self-generation of diverse data, self-evaluation for multi-objective preference data,self-alignment for multi-objective optimization and self-iteration for further improvement. Our approach allows for generating varied reports tailored to specific clinical objectives, enhancing the overall performance of the RRG model iteratively. Unlike existing methods, our frame-work significantly increases data quality and optimizes performance through iterative multi-objective optimization. Experimental results demonstrate that our method surpasses previous approaches, achieving state-of-the-art performance across multiple evaluation metrics.
title Online Iterative Self-Alignment for Radiology Report Generation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2505.11983