ROI-Aware ResNet for Retinal OCT Classification

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The accurate and timely classification of retinal OCT scans remains complex due to challenges in Region of Interest (ROI) identification, variability in manual interpretation, and the time-consuming nature of manual analysis, which can lead to misdiagnoses and delayed treatment.

Innovation Solution

A Region-of-Interest aware 2D Residual Neural Network (ResNet) is developed to automate and enhance the analysis of retinal OCT scans, focusing on efficient processing and accurate diagnosis by identifying and analyzing relevant regions within the scans.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual interpretation of OCT scans is performed, then diagnostic expertise can be applied, but processing time increases and consistency decreases

Engineering Contradiction:

Improvediagnostic accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system enables automated self-diagnosis of retinal conditions through the ROI-aware ResNet model, which automatically identifies regions of interest and classifies pathologies without requiring manual clinician intervention for each scan, thereby reducing processing time while maintaining diagnostic accuracy

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical process of manual image interpretation by clinicians with an automated neural network system. The ROI-aware ResNet model processes OCT scans algorithmically, substituting human visual analysis with machine learning-based automated detection and classification

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If manual analysis of OCT scans is performed, then expert judgment can be applied, but variability in assessments leads to misdiagnoses

Engineering Contradiction:

Improvediagnostic consistencyVSAvoidsystem complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The automated system performs consistent diagnostic assessments without human variability, applying the same computational rules and criteria to every scan processed, thereby eliminating inter-observer and intra-observer variability inherent in manual interpretation

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system transforms the diagnostic process from subjective human judgment to objective computational analysis by changing the parameters of assessment from clinician expertise and experience to standardized neural network feature extraction and classification criteria

Inventive Principle:

Principle #35Parameter changes

3Speed

If early detection is prioritized, then timely treatment can be enabled, but processing speed must increase

Engineering Contradiction:

Improveprocessing speedVSAvoiddetection accuracy

Core Design Contradiction:

SpeedVSMeasurement precision

Solution Approach 1:

The system extracts and focuses computational resources on only the relevant regions of interest within OCT scans rather than analyzing the entire image. By identifying and concentrating on pathological areas, the model achieves rapid processing without sacrificing detection accuracy

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the OCT scan analysis into two stages: first identifying regions of interest, then performing detailed classification only on those segments. This segmentation enables parallel processing and reduces the overall computational burden, increasing processing speed while maintaining precision

Inventive Principle:

Principle #1Segmentation

4Productivity

If comprehensive analysis of entire OCT scans is performed, then all regions are examined, but processing time increases significantly

Engineering Contradiction:

ImprovethroughputVSAvoidanalysis time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system extracts and isolates only the diagnostically relevant regions of interest from the full OCT scan, discarding or skipping analysis of irrelevant areas. This extraction approach maintains high throughput by focusing computational effort where it is most needed

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system performs partial analysis by examining only the necessary portions of the OCT scan (regions of interest) rather than conducting exhaustive analysis of the entire image, achieving sufficient diagnostic accuracy with reduced processing time

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20240281968A1System and method for retinal optical coherence tomography classification using region-of-interest aware resnet

Publication Date: 2024.08.22 MG HEALTH TECH LLC

AI summary

A retinal optical coherence tomography (OCT) image analysis (ROCTIA) system (1200) and method for analyzing one or more retinal scan images of an eye of a user to identify one or more retinal conditions of the eye. The system includes a scanner device (1202) and a processor (1204) configured with a Region-of-Interest Aware (ROI-Aware) Residual Network (ResNet). The scanner device (1202) configured to scan the eye of the user to obtain the one or more retinal scan images of the eye of the user. The processor (1204) classifies each of the one or more retinal scan images based on a region of interest (ROI) in each of the one or more retinal scan images. The ROI is obtained in real-time while the one or more retinal scan images are obtained. The processor (1204) identifies one or more retinal conditions of the eye based on one or more retinal scan images.