Abstract
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 while only referring 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.
Highlights
In recent years, deep neural networks (DNNs) have gained tremendous attention and shown outstanding performances in many different computer vision tasks
We showed that we can compute informative, interpretable uncertainty estimates for the skin lesion diagnosis task using the connection between the dropout operation and approximate Bayesian inference [39,51]
We present an approximate risk-aware deep Bayesian model, named Bayesian DenseNet-169, which outputs an estimate of the model uncertainty with no additional parameter or major change in the network’s architecture
Summary
Deep neural networks (DNNs) have gained tremendous attention and shown outstanding performances in many different computer vision tasks. These models are composed of stacks of processing layers to learn powerful representations from high-dimensional input data with multiple levels of abstraction [1]. Deep networks have even matched or surpassed human-level performance in tasks such as diabetic retinopathy detection [6] and skin lesion classification [7] Such systems can be employed to detect patients at risk from a prescreening examination, considerably decrease the physicians’ workload and diagnostic errors. As a result, optimizing the quality of the interaction between physicians and CAD systems as a team is often overlooked
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
