BINOCULAR KERATOCONUS DIAGNOSIS METHOD BASED ON MULTI-MODAL DATA

A binocular keratoconus diagnosis method based on multi-modal data. Four corneal dioptric maps and corneal absolute height data of two eyes are used; the correlation between two eyes is taken into consideration; a method based on a deep convolutional network, a traditional machine learning method, that is, an svm method, and an adjustable BSF height map enhancement method are combined to identify the sensitivity and specificity of a lesion, and the sensitivity and specificity thereof are balanced; and the morbidity of keratoconus using a patient as a unit is comprehensively determined from a plurality of dimensions, such that the diagnosis method has stronger robustness and accuracy. Procédé de diagnostic de kératocône binoculaire basé sur des données multimodales. Quatre cartes dioptriques cornéennes et des données de hauteur absolue cornéenne de deux yeux sont utilisées ; la corrélation entre deux yeux est prise en considération ; un procédé basé sur un réseau convolutif profond, un procédé d'apprentissage