从一组视觉变压器中提炼知识,以改善乳房超声波的分类
1Weill Cornell Medicine, New York, NY 10021.
Academic radiology
|September 4, 2023
概括
这项研究开发了一种深度学习模型,用于乳房超声波图像分类. 综合学习和知识蒸显著提高了诊断准确性,优于单个模型.
科学领域:
- 医疗成像医学成像
- 人工智能的人工智能
- 机器学习 机器学习
背景情况:
- 精确的乳房超声波图像的分类对于早期癌症检测至关重要.
- 深度学习模型为医疗图像的自动化分析提供了潜力.
研究的目的:
- 开发和评估一种深度学习模型,用于将乳房超声波图像分类为良性或恶性.
- 探索视觉转换器 (ViT),集合学习和知识蒸的有效性.
主要方法:
- 使用乳房超声波图像 (BUSI) 数据集进行培训和验证.
- 将ViT的性能与卷积神经网络 (CNN) 的性能进行比较.
- 调查监督,自我监督和随机初始化的ViT模型,其次是集合和知识蒸技术.
主要成果:
- 视觉转换器 (ViT) 在乳房超声波图像分类中表现优于CNN.
- 自主监督的ViT模型与监督和随机初始化模型相比,表现优越.
- 一组ViT模型实现了0.977的接收器操作特征曲线下的面积 (AuROC) 和0.965的精度回调曲线下的面积 (AuPRC),优于单个模型的性能. 一个通过知识蒸培训的ViT获得了类似的结果 (AuROC 0.972,AuPRC 0.960).
结论:
- 转移学习和集体学习可以结合起来,以提高乳房超声波分类深度学习模型的性能.
- 知识蒸使单个ViT模型能够实现与一组模型相美的性能,提供更有效的解决方案.
相关概念视频
Imaging Studies II: Ultrasonography
28
IntroductionUltrasonography, or renal ultrasound, is a noninvasive medical imaging technique that uses high-frequency sound waves to visualize the kidneys, ureters, bladder, and surrounding tissues.Indications for Urinary System UltrasonographyUrinary system ultrasonography is indicated in various clinical scenarios, such as:Kidney Stones (Urolithiasis): To detect and monitor the size and presence of kidney or urinary tract stones.Hydronephrosis: To assess the dilation of the renal pelvis and...
28
Ultrasonography
4.5K
Ultrasonography is an imaging technique that uses high-frequency sound waves to visualize the body's internal structures. It is a non-invasive and safe procedure that does not involve the use of ionizing radiation, making it widely used in various medical fields. Ultrasonography is used to study heart function, blood flow in the neck or extremities, certain conditions such as gallbladder disease, and fetal growth and development.
During an ultrasonography procedure, a handheld device called...
During an ultrasonography procedure, a handheld device called...
4.5K


