Exploring Artificial Intelligence Methods for the Automatic Measurement of a New Biomarker Aiming at Glaucoma Diagnosis

• Gabriel C. Fernandes UNISINOS
• Fabio Lavinsky UNISINOS
• Sandro José Rigo UNISINOS
• Henrique C. Bohn UNISINOS

Resumo

Using technologies capable of providing retina structure high-resolution images is one of the most widespread means of identifying structural changes that may indicate the onset or progression of visual impairment. Automated glaucoma detection using optical coherence tomography is still considered an area needing further research. Several manual analyzes are currently performed over the generated by imaging equipment. This work presents an approach to foster automatic glaucoma evaluation considering convolutional neural networks for semantic segmentation of retinal layers through optical coherence tomography images and image processing for measuring the cup region in the optic nerve head portion. We provide a quantitative evaluation comparing the results obtained by a specialist physician. The work’s main contribution consists of the first approach supporting the automation of a new biomarker for diagnosing glaucoma.