Researchers from the Johns Hopkins Applied Physics Laboratory and collaborators at the Johns Hopkins School of Medicine have developed image analysis and machine learning tools to automate the detection of age-related macular degeneration (AMD). AMD is a degenerative disease of the central portion of the retina (the macula) and a leading cause of loss of central vision which is required for activities of daily living, including reading, driving and watching television.
The developed algorithm, based on deep convolutional neural networks, was compared with human graders in several experiments evaluating over 130,000 images from 4613 patients to distinguish disease-free/early stages from the referable intermediate/advanced stages of AMD. The deep convolutional neural network method yielded accuracy (SD) that ranged between 88.4% (0.5%) and 91.6% (0.1%), the area under the receiver operating characteristic curve was between 0.94 and 0.96, a result comparable with a gold standard included in the National Institutes of Health Age-related Eye Disease Study data set.
Early identification of AMD can be time-intensive and requires expertly trained individuals. This study suggests that automated algorithms could address the costs of screening or monitoring, access to health care, and the assessment of novel treatments that address the development or progression of AMD.
Read more here: Neural networks detect AMD