CAD-PsorNet:深度转移学习用于皮肤牛皮的计算机辅助诊断
Chandan Chakraborty1, Unmesh Achar2, Sumit Nayek1
1National Institute of Technical Teachers' Training & Research (Deemed to be University), Kolkata, 700106, India.
Scientific reports
|November 3, 2024
概括
一个自动化的深度学习框架从皮肤图像中准确地检测出牛皮. 移动NetV1实现了99.13%的准确性,改善了这种慢性皮肤疾病的诊断.
科学领域:
- 皮肤病学和人工智能的人工智能
- 医学图像分析 医学图像分析
- 计算生物学 计算生物学
背景情况:
- 牛皮是一种慢性,炎症性皮肤疾病,由于其多样化的呈现,具有诊断挑战.
- 准确及时诊断对于有效管理牛皮至关重要.
- 北印度成年人口中牛皮的患病率从0.44%到2.8%不等.
研究的目的:
- 开发和评估使用深度转移学习检测牛皮的自动化框架.
- 为了比较不同深度学习模型在牛皮图像识别方面的性能.
- 通过超参数调整优化开发的模型,并评估其诊断准确性.
主要方法:
- 收集了325张原始牛皮图像的数据集,并将其处理成496张图像补丁.
- 四个深度转移学习模型 (VGG16,VGG19,MobileNetV1,ResNet-50) 用于特征提取和分类.
- 模型适应了密集,掉落和输出层;采用了超参数调整和AdaGrad优化器.
主要成果:
- 移动NetV1最初显示了94.84%的灵敏度,89.37%的特异性和97.24%的准确性.
- 在超参数调整后,该方法获得了94.25%的灵敏度,96.42%的特异性和99.13%的整体准确度.
- 该模型的性能,以0.98的子系数,超过了非机器学习方法.
结论:
- 开发的自动化牛皮图像识别框架显示了高准确性和有效性.
- 深度转移学习,特别是MobileNetV1,显示出改善牛皮诊断的巨大潜力.
- 未来的工作需要多样化的数据集,以提高跨不同人口和牛皮变异的模型稳定性.
相关概念视频
01:30Skin Cancer
3.8K
Skin cancer is a type of cancer that occurs when there is an abnormal growth of skin cells, usually triggered by damage to the DNA within the skin cells. It is primarily caused by exposure to ultraviolet (UV) radiation from the sun or artificial sources like tanning beds. Skin cancer is the most common type of cancer worldwide, and its incidence continues to rise.
Basal Cell Carcinoma (BCC): BCC is the most common type of skin cancer, accounting for about 80% of cases. It typically develops in...
Basal Cell Carcinoma (BCC): BCC is the most common type of skin cancer, accounting for about 80% of cases. It typically develops in...
3.8K
01:22iPS Cell Differentiation
2.6K
The ability of induced pluripotent stem cells or iPSCs to differentiate into most body cell types has stimulated repair and regenerative medicine research over the past few decades. iPSC-derived blood cells, hepatocytes, beta islet cells, cardiomyocytes, neurons, and other cell types can repair injuries or regenerate damaged tissue in diseases such as diabetes and neurodegenerative disorders.
2.6K