Ophthalmologic Image Processing with Confidence-Based Attention Areas

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing ophthalmologic image processing systems using machine learning algorithms for disease and structure analysis in eyes face variability in certainty and difficulty in efficiently presenting multiple analysis results to users, leading to inadequate user assistance in operations like diagnosis.

Innovation Solution

An ophthalmologic image processing device and program that utilize a mathematical model trained by a machine learning algorithm to acquire analysis results, generate supplemental distribution information indicating weight and certainty maps, and display attention areas based on these maps to assist users in diagnosing eye diseases and structures.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If machine learning algorithms are used to analyze ophthalmologic images for disease and structure, then analysis results can be obtained, but the certainty of the analysis varies depending on the subject eye

Engineering Contradiction:
Improveanalysis certaintyVSAvoidanalysis reliability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent introduces a confidence map as an intermediary element that bridges the analysis results and the original image. This confidence map visually represents the certainty level at different regions of the eye image, allowing users to understand the reliability of analysis results without directly exposing the variability in machine learning predictions. The confidence map serves as a mediator that translates abstract algorithmic uncertainty into visual information.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent employs color changes to represent different confidence levels in the confidence map. Regions with high confidence are displayed in one color (e.g., bright or saturated), while regions with low confidence are displayed in another color (e.g., dark or desaturated). This visual encoding allows users to quickly assess the reliability of analysis results across different areas of the eye image without requiring numerical interpretation.

Inventive Principle:
Principle #32Color changes

2Loss of information

If multiple analysis results are provided to users, then comprehensive information is available, but it becomes difficult for users to recognize and process the results efficiently

Engineering Contradiction:
Improveinformation completenessVSAvoidresult processing efficiency
Core Design Contradiction:
Loss of informationVSEase of operation

Solution Approach 1:

The patent segments the complex analysis results into multiple visual components: the original ophthalmologic image, the confidence map, and the analysis results display. By dividing the information presentation into separate, manageable segments, users can process each component systematically rather than being overwhelmed by a single complex result display. The confidence map itself is segmented into regional zones based on certainty levels.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent adds a new dimensional layer by introducing the confidence map as a separate visual dimension alongside the original image and analysis results. This third dimension allows users to simultaneously view the anatomical structure, the analysis findings, and the confidence levels without cluttering any single display. The confidence map can be overlaid on the original image or displayed as a separate panel, providing spatial context for the analysis results.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP3690888B1Ophthalmologic image processing device and ophthalmologic image processing program
Publication Date: 2025.07.02 NIDEK CO LTD
  • EP3690888B1 patent drawingFigure 1
  • EP3690888B1 patent drawingFigure 2
  • EP3690888B1 patent drawingFigure 3

AI summary

A processor (23) of an ophthalmologic image processing device (21) acquires an ophthalmologic image photographed by an ophthalmologic image photographing device (11). The processor inputs the ophthalmologic image into a mathematical model trained by a machine learning algorithm to acquire a result of an analysis relating to at least one of a specific disease and a specific structure of a subject eye. The processor acquires information of a distribution of weight relating to an analysis by a mathematical model, as supplemental distribution information, for which an image area of the ophthalmologic image input into the mathematical model is set as a variable. The processor sets a part of the image area of the ophthalmologic image, as an attention area (48), based on the supplemental distribution information. The processor acquires an image of a tissue including the attention area among a tissue of the subject eye and displays the image on a display unit (28).