Diabetic Peripheral Neuropathy Risk Assessment using Digital Fundus Photographs and Machine Learning
Jeremy Benson, Trilce Estrada, Mark R. Burge, Peter Solíz
Abstract
In this work, we demonstrate a novel approach to assessing the risk of Diabetic Peripheral Neuropathy (DPN) using only the retinal images of the patients. Our methodology consists of convolutional neural network feature extraction, dimensionality reduction and feature selection with random projections, combination of image features to case-level representations, and the training and testing of a support vector machine classifier. Using clinical diagnosis as ground truth for DPN, we achieve an overall accuracy of 89% on a held-out test set, with sensitivity reaching 78% and specificity reaching 95%.
Topics & Concepts
Convolutional neural networkPeripheral neuropathyArtificial intelligenceComputer scienceFeature extractionFeature selectionTest setGround truthPattern recognition (psychology)Dimensionality reductionSupport vector machineRandom forestClassifier (UML)Computer visionMedicineDiabetes mellitusEndocrinologyRetinal Imaging and AnalysisRetinal Diseases and TreatmentsGlaucoma and retinal disorders