Light-Invariant Pupil Reactivity Diagnosis via ML

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for assessing pupil reactivity in ophthalmology are subjective and prone to errors due to variations in external light conditions, leading to inconsistent and unreliable measurements.

Innovation Solution

A machine learning model is trained using light-corrected pupillometry parameters to construct a Neuro-Pupillary Index (NPx) that is invariant to external light conditions, utilizing a combination of static and dynamic pupillometric parameters, and light sensor data to provide a standardized numerical scale for pupil reactivity assessment.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If manual pupil measurement methods are used, then the assessment process is simple and quick, but the measurement precision and reliability deteriorate due to subjective errors and lack of standardization

Engineering Contradiction:

Improvesimplicity of assessment processVSAvoidreliability of pupil reactivity measurement

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The patent replaces manual mechanical measurement methods with an automated digital imaging system that captures pupil responses using a camera or smartphone. The system automatically extracts pupillometry parameters (pupil diameter, constriction velocity, dilation velocity) from images and computes the NPx index, eliminating subjective human error while maintaining operational simplicity through automated processing algorithms.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Device complexity

If traditional pupillometry parameters are used without light correction, then the measurement process is straightforward, but the reliability deteriorates due to variability introduced by external light conditions

Engineering Contradiction:

Improvesimplicity of measurement processVSAvoidconsistency of pupil reactivity assessment

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The patent introduces light condition parameters (ambient light intensity, camera exposure settings ISO and shutter speed) into the measurement process. The system captures these parameters simultaneously with pupillometry data and uses them to correct the extracted parameters through a light correction model, thereby compensating for external light variability and improving measurement reliability without significantly increasing operational complexity.

Inventive Principle:

Principle #35Parameter changes

3Measurement precision

If light-corrected pupillometry parameters and machine learning models are implemented, then the reliability and accuracy of pupil reactivity assessment improve, but the device complexity and computational requirements increase

Engineering Contradiction:

Improveaccuracy of pupil reactivity diagnosisVSAvoidcomplexity of processing system

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent pre-trains machine learning models using large datasets of pupillometry measurements with corresponding light conditions and clinical outcomes. The trained models (including light correction models and NPx index computation models) are stored in the system, allowing rapid inference during actual measurements without requiring complex real-time training. This approach achieves high diagnostic accuracy while keeping the operational system relatively simple.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent introduces a light correction model as an intermediary component between raw pupillometry parameter extraction and final NPx index computation. This intermediary model corrects the extracted parameters for light condition effects before they are fed into the diagnostic algorithm, thereby improving accuracy without requiring the final diagnostic model to directly process complex light variability.

Inventive Principle:

Principle #24Intermediary (Mediator)

4Reliability

If standardized numerical scales for pupil reactivity are implemented, then the diagnostic consistency improves across different settings, but the difficulty of detecting and measuring increases due to the need for multiple parameter measurements and corrections

Engineering Contradiction:

Improvestandardization of diagnostic resultsVSAvoidcomplexity of parameter extraction and correction

Core Design Contradiction:

ReliabilityVSDifficulty of detecting and measuring

Solution Approach 1:

The patent combines multiple pupillometry parameters (pupil diameter, constriction velocity, dilation velocity, reaction time) and light condition parameters (ambient light intensity, camera exposure settings) into a single standardized NPx index value. This merging process integrates all measurements and corrections into one comprehensive diagnostic metric, achieving standardization and consistency while simplifying the final output interpretation.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentEP4574018A1A method for training a machine learning model for automatically diagnosing abnormal pupil behaviour based on assessing pupil reactivity invulnerable to external light conditions, a computer program comprising instructions thereof, and a system implementing the model

Publication Date: 2025.06.25 SOLVEMED GRP SP ZOO

Figure 1
Figure 2(a)~2(b)
Figure 3

AI summary

A present inventions describe a method for training a machine learning model for automatically diagnosing abnormal pupil behaviour based on assessing pupil reactivity invulnerable to external light conditions, a computer program comprising instructions thereof, and a system implementing the model.