2020/09/09 by Gabriel García, Adrián Colomer, García, Gabriel +4
Medicine · #FOS: Electrical engineering #Glaucoma and retinal disorders #Image and Video Processing (eess.IV) #Retinal Diseases and Treatments #Retinal Imaging and Analysis #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2009.04190
openalex publication_date 2020/09/09 · openalex created_date 2022/10/02 · openalex updated_date 2026/07/28
Taking into account that glaucoma is the leading cause of blindness\nworldwide, we propose in this paper three different learning methodologies for\nglaucoma detection in order to elucidate that traditional machine-learning\ntechniques could outperform deep-learning algorithms, especially when the image\ndata set is small. The experiments were performed on a private database\ncomposed of 194 glaucomatous and 198 normal B-scans diagnosed by expert\nophthalmologists. As a novelty, we only considered raw circumpapillary OCT\nimages to build the predictive models, without using other expensive tests such\nas visual field and intraocular pressure measures. The results ratify that the\nproposed hand-driven learning model, based on novel descriptors, outperforms\nthe automatic learning. Additionally, the hybrid approach consisting of a\ncombination of both strategies reports the best performance, with an area under\nthe ROC curve of 0.85 and an accuracy of 0.82 during the prediction stage.\n