MedCLM: Learning to Localize and Reason via a CoT-Curriculum in Medical Vision-Language Models

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Kim, Soo Yong, Cho, Suin, Yun, Vincent-Daniel, Hwang, Gyeongyeon
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908577378074624
author Kim, Soo Yong
Cho, Suin
Yun, Vincent-Daniel
Hwang, Gyeongyeon
author_facet Kim, Soo Yong
Cho, Suin
Yun, Vincent-Daniel
Hwang, Gyeongyeon
contents Bridging clinical diagnostic reasoning with AI remains a central challenge in medical imaging. We introduce MedCLM, an automated pipeline that converts detection datasets into large-scale medical visual question answering (VQA) data with Chain-of-Thought (CoT) reasoning by linking lesion boxes to organ segmentation and structured rationales. These contextual signals enable medical vision-language models to generate question-answer pairs with step-by-step reasoning. To utilize this data effectively, we propose an Integrated CoT-Curriculum Strategy composed of an Easy stage with explicit lesion boxes for visual grounding, a Medium stage that encourages implicit localization, and a Hard stage for weakly supervised reasoning. Experimental results demonstrate that MedCLM attains state-of-the-art performance on several medical VQA benchmarks, providing a scalable framework for developing clinically aligned medical vision-language models.
format Preprint
id arxiv_https___arxiv_org_abs_2510_04477
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MedCLM: Learning to Localize and Reason via a CoT-Curriculum in Medical Vision-Language Models
Kim, Soo Yong
Cho, Suin
Yun, Vincent-Daniel
Hwang, Gyeongyeon
Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
Machine Learning
Bridging clinical diagnostic reasoning with AI remains a central challenge in medical imaging. We introduce MedCLM, an automated pipeline that converts detection datasets into large-scale medical visual question answering (VQA) data with Chain-of-Thought (CoT) reasoning by linking lesion boxes to organ segmentation and structured rationales. These contextual signals enable medical vision-language models to generate question-answer pairs with step-by-step reasoning. To utilize this data effectively, we propose an Integrated CoT-Curriculum Strategy composed of an Easy stage with explicit lesion boxes for visual grounding, a Medium stage that encourages implicit localization, and a Hard stage for weakly supervised reasoning. Experimental results demonstrate that MedCLM attains state-of-the-art performance on several medical VQA benchmarks, providing a scalable framework for developing clinically aligned medical vision-language models.
title MedCLM: Learning to Localize and Reason via a CoT-Curriculum in Medical Vision-Language Models
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2510.04477