基于乳腺组织结构的乳腺癌分类,使用自主监督学习中的拼图任务
Keisuke Sugawara1, Eichi Takaya2,3, Ryusei Inamori4
1Department of Diagnostic Radiology, Tohoku University Graduate School of Medicine, 2-1 Seiryo-machi, Aoba-ku, Sendai, Miyagi, 980-8575, Japan.
使用拼图任务进行自我监督学习,有效地表征乳腺组织结构,用于乳房摄影中的癌症分类. 这种方法显示了提高诊断准确性的潜力,特别是当数据有限时.
科学领域:
- 医疗成像医学成像
- 人工智能的人工智能
- 在瘤学瘤学.
背景情况:
- 自主监督学习 (SSL) 使用未标记的数据进行医学深度学习.
- 在SSL中的拼图任务学习图像特征和空间关系.
- 目前用于乳腺癌诊断的深度学习模型缺乏人类放射科医生的综合方法.
研究的目的:
- 为了评估拼图任务在特征乳腺组织的乳腺分类的有效性.
- 将SSL Jigsaw预训练与其他乳腺癌检测方法进行比较.
主要方法:
- 在中国乳房学数据库 (CMMD) 上比较了四个预训练管道:IN-Jig,Scratch-Jig,IN和Scratch.
- 模型对二元乳腺癌分类进行了微调.
- 性能指标包括AUC,灵敏度和特异性,并使用Grad-CAM进行可视化.
主要成果:
- 基于任务的拼图模型 (IN-Jig和Scratch-Jig) 实现了高AUC (0.925和0.921).
- 所有车型都表现出强的表现,而Jigsaw预训练显示出了竞争力的结果.
- 详细分析显示,在不同放射学发现和乳腺密度之间,性能差异不同.
结论:
- 拼图拼图任务是乳腺癌分类的有价值的SSL预训练方法.
- 这种方法可以提高乳房造影的诊断准确性,特别是在有限的标记数据的情况下.
更多相关视频
15:48Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
Published on: December 15, 2014
08:59Morphology-Based Distinction Between Healthy and Pathological Cells Utilizing Fourier Transforms and Self-Organizing Maps
Published on: October 28, 2018
相关概念视频
Mouse Models of Cancer Study
The development of transgenic, knockout, and knock-in mice has led to an exponential increase in their use as model organisms in research,...
Mouse Models of Cancer Study
The development of transgenic, knockout, and knock-in mice has led to an exponential increase in their use as model organisms in research,...
Classification of Epithelial Tissues: Overview
Based on the number of cell layers,...
Classification of Epithelial Tissues: Stratified Epithelium
Classification of Connective Tissues
Connective Tissue Proper
Connective tissue proper is the most abundant class of connective tissues. As its name implies, it predominantly connects different tissues in the body. Depending on the cell types, ground substance, viscosity, and fiber types in the ECM, connective tissue proper is further categorized into loose and dense.
Classification of Systems-I
Homogeneity dictates that if an input x(t) is multiplied by a constant c, the output y(t) is multiplied by the same constant. Mathematically, this is expressed as:
