MRGen:基於擴散的可控資料引擎,用於 MRI 分割向未標記的模態靠攏

MRGen: Diffusion-based Controllable Data Engine for MRI Segmentation towards Unannotated Modalities

December 4, 2024
作者: Haoning Wu, Ziheng Zhao, Ya Zhang, Weidi Xie, Yanfeng Wang
cs.AI

摘要

最近,深度神經網絡在醫學影像分割方面取得了令人印象深刻的進展,然而異構模態和標記稀缺限制了在未標記模態上發展分割模型。本文探討了在醫學應用中利用生成模型的新範式:可控合成未標記模態的數據,而無需註冊數據對。具體而言,本文在以下方面做出了貢獻:(i)我們收集並整理了一個大規模放射學影像-文本數據集MedGen-1M,包括模態標籤、屬性、區域和器官信息,以及部分器官遮罩標註,以支持可控醫學影像生成的研究;(ii)我們提出了一種基於擴散的數據引擎,稱為MRGen,它可以根據文本提示和遮罩進行生成,合成缺乏遮罩標註的多樣模態的MR影像,以訓練未標記模態上的分割模型;(iii)我們在各種模態上進行了廣泛實驗,說明我們的數據引擎能夠有效合成訓練樣本,並將MRI分割擴展到未標記模態。
English
Medical image segmentation has recently demonstrated impressive progress with deep neural networks, yet the heterogeneous modalities and scarcity of mask annotations limit the development of segmentation models on unannotated modalities. This paper investigates a new paradigm for leveraging generative models in medical applications: controllably synthesizing data for unannotated modalities, without requiring registered data pairs. Specifically, we make the following contributions in this paper: (i) we collect and curate a large-scale radiology image-text dataset, MedGen-1M, comprising modality labels, attributes, region, and organ information, along with a subset of organ mask annotations, to support research in controllable medical image generation; (ii) we propose a diffusion-based data engine, termed MRGen, which enables generation conditioned on text prompts and masks, synthesizing MR images for diverse modalities lacking mask annotations, to train segmentation models on unannotated modalities; (iii) we conduct extensive experiments across various modalities, illustrating that our data engine can effectively synthesize training samples and extend MRI segmentation towards unannotated modalities.

Summary

AI-Generated Summary

PDF52December 6, 2024