Synthetic Tumors Make AI Segment Tumors Better Qixin Hu1Junfei Xiao2Yixiong Chen3Shuwen Sun4Jie-Neng Chen2 Alan Yuille2Zongwei Zhou2

2025-05-02 0 0 4.09MB 5 页 10玖币

侵权投诉

Synthetic Tumors Make AI Segment Tumors Better

Qixin Hu1Junfei Xiao2Yixiong Chen3Shuwen Sun4Jie-Neng Chen2

Alan Yuille2Zongwei Zhou2,∗

1Huazhong University of Science and Technology 2Johns Hopkins University

3Fudan University 4The First Afﬁliated Hospital of Nanjing Medical University

Code and Visual Turing Test: https://github.com/MrGiovanni/SyntheticTumors

Abstract

We develop a novel strategy to generate synthetic tumors. Unlike existing works,

the tumors generated by our strategy have two intriguing advantages: (1) realistic in

shape and texture, which even medical professionals can confuse with real tumors;

(2) effective for AI model training, which can perform liver tumor segmentation

similarly to a model trained on real tumors—this result is unprecedented because

no existing work, using synthetic tumors only, has thus far reached a similar or even

close performance to the model trained on real tumors. This result also implies

that manual efforts for developing per-voxel annotation of tumors (which took

years to create) can be considerably reduced for training AI models in the future.

Moreover, our synthetic tumors have the potential to improve the success rate of

small tumor detection by automatically generating enormous examples of small (or

tiny) synthetic tumors.

1 Introduction

Artiﬁcial intelligence (AI) has dominated medical image segmentation [

], but training an AI

model (

e.g.,

U-Net [

]) often requires a large number of detailed per-voxel annotations. Annotating

medical images is not only expensive and time-consuming, but also requires extensive medical

expertise, and sometimes needs the assistance of radiology reports and biopsy results to precisely

annotate a tumor [

]. Due to its high annotation cost, only a total of roughly 100 CT scans with

annotated liver tumors are publicly available (provided by LiTS [

]) for training and testing models.

Generating synthetic tumors is an attractive research topic. There are some early attempts at generating

COVID-19 infections on Chest CT scans [

], abdominal tumors in CT scans [

], diabetic lesions

on retinal images [

], brain tumors on MRI images [

], and cancers in ﬂuorescence microscopy

images [

]. However, the synthetic tumors in those existing studies appear very different from the real

tumors, and AI models trained using synthetic tumors perform signiﬁcantly worse than those trained

using real tumors due to the pronounced domain gap between real and synthetic tumors. What makes

synthesizing tumors so hard? There are several important factors: shape, intensity, size, location, and

most importantly, texture. In this paper, we develop a hand-crafted heuristic strategy to synthesize

liver tumors in abdominal CT scans. Our synthetic tumors are realistic—even medical professionals

can confuse them with real tumors in the Visual Turing Test [

] (Figure 1A). Besides, AI models

trained on our synthetic tumors can segment real tumors similar to those trained on real tumors

with expensive, detailed per-voxel annotation. As shown in Figure 1B, the model trained on our

(label-free) synthetic tumors achieves a Dice Similarity Coefﬁcient (DSC) of 52.0% for segmenting

real liver tumors, whereas AI trained on real tumors obtains a DSC of 52.3% (no statistical difference

between the two performances). These results are unprecedented because no existing work, using

synthetic tumors only, has thus far reached a similar performance to the model trained on real tumors.

∗Corresponding author: Zongwei Zhou (zzhou82@jh.edu)

36th Conference on Neural Information Processing Systems (NeurIPS 2022).

arXiv:2210.14845v1 [eess.IV] 26 Oct 2022

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

SyntheticTumorsMakeAISegmentTumorsBetterQixinHu1JunfeiXiao2YixiongChen3ShuwenSun4Jie-NengChen2AlanYuille2ZongweiZhou2,1HuazhongUniversityofScienceandTechnology2JohnsHopkinsUniversity3FudanUniversity4TheFirstAfliatedHospitalofNanjingMedicalUniversityCodeandVisualTuringTest:https://github.com/MrGiov...

展开>> 收起<<

Synthetic Tumors Make AI Segment Tumors Better Qixin Hu1Junfei Xiao2Yixiong Chen3Shuwen Sun4Jie-Neng Chen2 Alan Yuille2Zongwei Zhou2.pdf

共5页,预览1页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Synthetic Tumors Make AI Segment Tumors Better Qixin Hu1Junfei Xiao2Yixiong Chen3Shuwen Sun4Jie-Neng Chen2 Alan Yuille2Zongwei Zhou2

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: