OCT Glaucoma Estimation Using Multi-Orientation Cross-Section Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting glaucoma lack precision and accuracy in early detection, necessitating improved techniques for timely intervention.
Innovation Solution
An ophthalmic information processing apparatus that generates multiple images with different cross-sectional orientations from three-dimensional OCT data using machine learning, employing a glaucoma estimation model to accurately classify eyes as glaucoma or normal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing methods for detecting glaucoma are used, then the detection process is simple, but the precision and accuracy in early detection are insufficient
Solution Approach 1:
The patent segments the glaucoma detection process into multiple specialized learned models, each trained on specific image types (fundus images, OCT images, anterior segment images). This segmentation allows each model to specialize in detecting specific glaucoma indicators from its designated image type, thereby improving overall detection precision while maintaining manageable system complexity through modular architecture
Solution Approach 2:
The patent transitions from two-dimensional image analysis to three-dimensional volumetric analysis by incorporating OCT (optical coherence tomography) images that provide depth information. This dimensional expansion enables detection of subsurface structural changes in the eye that are invisible in traditional 2D fundus images, significantly improving early glaucoma detection accuracy
2Reliability
If existing methods for detecting glaucoma are used, then the system is simple to operate, but the accuracy for timely intervention is insufficient
Solution Approach 1:
The patent performs preliminary actions by pre-training multiple specialized learned models on large datasets of various eye images before actual glaucoma screening. These pre-trained models are ready to immediately analyze patient images, enabling rapid and accurate detection without requiring time-consuming manual analysis or model training during the screening process itself
Solution Approach 2:
The patent creates multiple copies of learned models, each specialized for different image types (fundus, OCT, anterior segment). These model copies work in parallel to analyze different aspects of eye health simultaneously, improving detection accuracy through comprehensive multi-modal analysis while maintaining efficient processing speed through parallel computation
Data Source
AI summary
An ophthalmic information processing apparatus includes an acquisition unit and a disease estimation unit. The acquisition unit is configured to acquire a plurality of images with different cross-sectional orientations from each other of a subject's eye. The disease estimation unit is configured to output estimation information for estimating whether or not the subject's eye is a glaucoma eye from the images, using a plurality of learned models obtained by performing machine learning for each type of the images.