Risk-Aware Machine Learning Classifier for Skin Lesion Diagnosis
Risk-Aware Machine Learning Classifier for Skin Lesion Diagnosis
复制标题
DOI：
10.3390/jcm8081241
复制
发表时间：
2019-08-01
影响因子：
3.9
通讯作者：
Nguyen, Hien Van
中科院分区：
文献类型：
--
作者：
Mobiny, Aryan;Singh, Aditi;Nguyen, Hien Van
在安全性至关重要的医疗领域，了解机器学习系统何时对其预测不自信至关重要。理想情况下，机器学习算法应该只在高度确定其能力时才进行预测，否则将病例提交给医生。在本文中，我们研究了贝叶斯深度学习如何提高机器医生团队在皮肤病变分类任务中的表现。我们使用了公开可用的HAM 10000数据集，其中包括来自七种常见皮肤病变类别的样本：黑色素瘤（MEL）、黑色素细胞痣（NV）、基底细胞癌（BCC）、光化性角化病和上皮内癌（AKIEC）、良性角化病（BKL）、皮肤纤维瘤（DF）和血管（VASC）病变。我们的实验结果表明，贝叶斯深度网络可以将标准DenseNet-169模型的诊断性能从81.35%提高到83.59%，而不会产生额外的参数或繁重的计算。更重要的是，医生-机器混合工作流程的分类准确率达到90%，而只有35%的病例被推荐给医生。这些发现有望推广到其他医疗诊断应用。我们相信，风险感知机器学习方法的可用性将使机器学习技术在临床环境中得到更广泛的采用。
Knowing when a machine learning system is not confident about its prediction is crucial in medical domains where safety is critical. Ideally, a machine learning algorithm should make a prediction only when it is highly certain about its competency, and refer the case to physicians otherwise. In this paper, we investigate how Bayesian deep learning can improve the performance of the machine-physician team in the skin lesion classification task. We used the publicly available HAM10000 dataset, which includes samples from seven common skin lesion categories: Melanoma (MEL), Melanocytic Nevi (NV), Basal Cell Carcinoma (BCC), Actinic Keratoses and Intraepithelial Carcinoma (AKIEC), Benign Keratosis (BKL), Dermatofibroma (DF), and Vascular (VASC) lesions. Our experimental results show that Bayesian deep networks can boost the diagnostic performance of the standard DenseNet-169 model from 81.35% to 83.59% without incurring additional parameters or heavy computation. More importantly, a hybrid physician-machine workflow reaches a classification accuracy of 90% while only referring 35% of the cases to physicians. The findings are expected to generalize to other medical diagnosis applications. We believe that the availability of risk-aware machine learning methods will enable a wider adoption of machine learning technology in clinical settings.