轻量级深度学习模型用于对有机图像中的正常和异常血管系统的分类
Eunsu Yun1, Jongweon Kim1, Daesik Jeong2
1Department of Computer Science, Sangmyung University, Seoul 03016, Republic of Korea.
Sensors (Basel, Switzerland)
|January 10, 2026
概括
这项研究引入了一种快速,准确的深度学习模型,用于自动评估人类器官中的血管系统. 轻量级模型通过减少手动检查时间和主观性来确保可靠的实验结果.
科学领域:
- 生物医学工程 生物医学工程
- 计算生物学 计算生物学
- 细胞生物学 细胞生物学
背景情况:
- 人类有机体模型器官微环境用于研究.
- 在有机体中评估正常的血管形成对于实验可靠性至关重要.
- 目前的手动评估方法耗时且主观.
研究的目的:
- 开发一种轻量级的深度学习模型,用于在血管器官图像中自动分类正常和异常血管.
- 为了提高有机血管系统评估的效率和可重复性.
主要方法:
- 修改后的EfficientNet模型 (用ReLU取代SiLU,删除SE块) 用于图像分类.
- 该模型是通过来自共同培养实验的血管器官图像进行训练的.
- 数据增强和噪声添加被用来解决阶级不平衡.
主要成果:
- 拟议的修改3模型 (B0,B1,B2) 实现了高精度 (0.90-1.00).
- 在CPU上记录了51.1 FPS,36.0 FPS和32.4 FPS的实时推断速度.
- 与原始模型相比,观察到平均速度提高了70%.
结论:
- 开发的轻量级深度学习模型使有机体中有效和自动化的血管系统评估成为可能.
- 该框架提供了定量和可重现的分析,提高了基于有机体的研究的可靠性.
- 该模型的实时处理能力支持高吞吐量分析.
相关概念视频
Imaging Studies VII: Vascular Imaging
303
DefinitionRenal angiography, also known as renal arteriography, is an imaging technique used to obtain a comprehensive view of blood flow and the vascular structure of blood vessels in the kidneys and surrounding areas.PurposeRenal angiography detects blood vessel abnormalities in the kidneys, such as aneurysms, stenosis, thrombosis, vascular tumors, and renal artery stenosis. It evaluates kidney function and guides interventional treatments like angioplasty or stent placement.Pre-Procedure...
303
Anatomy of Blood Vessels
2.4K
The vascular system, an integral part of the circulatory system, comprises various blood vessels that play crucial roles in maintaining the body's homeostasis. These blood vessels form a complex and efficient circulatory network. The three primary categories of blood vessels are the arteries, veins, and capillaries.
Arteries
Arteries circulate oxygenated blood from the heart, except the pulmonary artery, which transports deoxygenated blood to the lungs. Large arteries, such as the aorta,...
Arteries
Arteries circulate oxygenated blood from the heart, except the pulmonary artery, which transports deoxygenated blood to the lungs. Large arteries, such as the aorta,...
2.4K


