医学图像检测和分割模型在肋骨骨折诊断中的应用进展  

A Review of Advances in Medical Image Detection and Segmentation Models for Rib Fracture Diagnosis

作　　者：赵婷婷[1] 郭宇 才佳 罗嗣频 孙志伟 畅卫功[1] ZHAO Tingting;GUO Yu;CAI Jia;LUO Sipin;SUN Zhiwei;CHANG Weigong(College of Artificial Intelligence,Tianjin University of Science and Technology,Tianjin 300457,China;Department of Radiology,Tianjin Hospital,Tianjin 300211,China)

机构地区：[1]天津科技大学人工智能学院,天津300457 [2]天津医院放射科,天津300211

出　　处：《天津科技大学学报》2025年第1期1-12,34,共13页Journal of Tianjin University of Science & Technology

基　　金：国家自然科学基金项目(61976156)。

摘　　要：肋骨骨折是一种高发疾病,具有较高的误诊率,对于医学图像处理技术的诊断速度和准确率要求较高,而图像检测与分割是实现肋骨骨折诊断的关键技术。本文综述了当前医学图像检测和分割模型在肋骨骨折诊断中的研究进展。首先对当前主流的图像检测和分割模型进行了系统介绍,然后对现有模型进行归纳和总结,分析了基于深度模型骨折诊断的各种算法,最后利用RibFrac挑战赛的公共数据库对模型进行对比分析。研究表明,基于图像检测和分割的肋骨骨折诊断方法在临床上已经有了长足的发展。然而,模型训练过程中容易受到样本数据的干扰,三维数据需要耗费大量的时间和资源,并且新模型和新技术难以实现。因此,在肋骨骨折诊断方面,图像检测和分割模型仍面临着许多挑战,有较大的发展空间。Rib fracture is a high incidence disease with high misdiagnosis rate,so the diagnostic speed and accuracy requirements for medical image processing technology are high,and image detection and segmentation is the key technology to achieve rib fracture diagnosis.Therefore,the current research progress in the application of medical image detection and segmentation models in rib fracture diagnosis is reviewed in this article.Specifically,the current mainstream image detection and segmentation models are firstly introduced systematically,then the present models are summarized and concluded,and various fracture diagnosis algorithms based on depth models are analysed,and finally,a comparative analysis of the models is conducted using the RibFrac Challenge public database.The study shows that rib fracture diagnosis methods based on image detection and segmentation have made great strides in clinical practice.However,the model training process is susceptible to interference from sample data,3D data is time-consuming and resource-intensive,and it is difficult to implement new models and technologies.Therefore,there are a lot of challenges and opportunities to develop in the diagnosis of rib fracture.

关 键 词：图像检测模型 图像分割模型 肋骨骨折 RibFrac U-Net 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]