Starting from a volumetric SD-OCT scan, the system generates a two-dimensional enface map of the optic nerve head by averaging and normalizing A-scan intensities. It then detects the disc boundary via a modified active contour framework: median filtering identifies low‐intensity regions, Canny edge mapping applies intensity‐weighted pixel costs, a Circular Hough transform estimates initial center and radius, and an energy-minimizing deformation refines the contour using smoothness, gradient orientation, and intensity terms. Cup delineation follows by segmenting the internal limiting membrane and retinal pigment epithelium layers, constructing a 3D surface, and intersecting that surface with an RPE‐based reference line or fixed‐distance plane. After radial smoothing and cubic interpolation, the system computes disc and cup areas, volumes, global and sectoral cup‐to‐disc and rim‐to‐disc ratios, plus vertical and horizontal C/D metrics. Finally, a statistical point distribution model trained on normal and abnormal shapes classifies the optic nerve head based on lower‐energy alignment.
Description
In this invention, a fully automated optic nerve head (ONH) assessment system is designed and developed based on spectral domain optical coherence tomography (SDOCT), with the aim to provide essential disc parameters for clinical analysis, early detection and monitoring of the progression for glaucoma and other optic nerve head diseases. The system can automatically provide subjective and reliable results of ONH evaluation. The input comprises from three dimensional (3D) OCT image taken from SDOCT.
Applications
Automated glaucoma screening software
Optic nerve head analytics platform
Ophthalmic imaging diagnostics integration
Teleophthalmology remote diagnosis solution
Clinical decision support system
Advantages
Fully automated, objective segmentation of optic disc and cup margins
Accurate quantification of ONH parameters (disc/cup area, volume, cup-to-disc and rim-to-disc ratios)
Robust performance against peripapillary atrophy and blood vessel occlusions
Early glaucoma detection and reliable monitoring of disease progression
Consistent, repeatable measurements derived from high-resolution 3D SD-OCT data
Automated normal vs. abnormal classification using trained shape models
IP Status
https://patents.google.com/patent/US7992999B2
