Skin Tone Calibration for Bias-Resistant Diagnostic Model Selection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing skin pigmentation classification systems, such as the Fitzpatrick scale, are inadequate for accurately defining skin color, leading to imprecise or missed diagnoses of skin conditions, particularly in non-white skin tones, due to biases and limitations in self-reporting and ethnic applicability.

Innovation Solution

An automated system using machine learning models to objectively analyze and classify skin tones by capturing and calibrating base skin tone images, selecting appropriate diagnostic models based on the patient's skin tone, and processing concern images for accurate diagnosis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If the Fitzpatrick scale is used for skin pigmentation classification, then the system is simple to operate, but the measurement precision is insufficient due to self-reporting errors and limited applicability to all skin types

Engineering Contradiction:
Improveease of operationVSAvoidmeasurement precision
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent replaces the manual self-reporting mechanism of the Fitzpatrick scale with an automated image-based classification system. A camera captures skin tone images, and machine learning models automatically determine skin type, eliminating human subjectivity and error while maintaining ease of use through automated processing

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent introduces an intermediary calibration system using color charts and reference standards between the raw image capture and final skin type classification. This intermediary step ensures accurate color representation across different devices and lighting conditions, improving measurement precision without complicating the user interface

Inventive Principle:
Principle #24Intermediary (Mediator)

2Device complexity

If a single machine learning model is used for all skin tones, then the device complexity is reduced, but the measurement precision deteriorates due to biases against non-white skin types

Engineering Contradiction:
Improvedevice complexityVSAvoidmeasurement precision
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent segments the skin tone classification task into multiple specialized machine learning models, each trained on specific skin tone ranges or ethnic groups. This segmentation allows each model to achieve high precision for its target population while the overall system maintains manageable complexity through modular architecture and automated model selection based on input image analysis

Inventive Principle:
Principle #1Segmentation

3Ease of operation

If skin diseases are examined only in commonly inspected areas, then the ease of operation is maintained, but the measurement precision of diagnosis deteriorates for skin cancers in non-UV exposed areas

Engineering Contradiction:
Improveease of operationVSAvoidmeasurement precision
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent implements a dynamic examination approach where the system automatically adjusts the inspection areas based on the patient's skin tone classification. For darker skin types, the system dynamically expands the examination to include non-UV exposed areas such as palms, soles, and mucous membranes, while maintaining ease of operation through automated guidance and adaptive imaging protocols

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3986251B1Using a set of machine learning diagnostic models to determine a diagnosis based on a skin tone of a patient
Publication Date: 2025.09.03 DIGITAL DIAGNOSTICS INC
  • EP3986251B1 patent drawingFigure 1
  • EP3986251B1 patent drawingFigure 2
  • EP3986251B1 patent drawingFigure 3

AI summary

Systems and methods are disclosed herein for determining a diagnosis based on a base skin tone of a patient. In an embodiment, the system receives a base skin tone image of a patient, generates a calibrated base skin tone image by calibrating the base skin tone image using a reference calibration profile, and determines a base skin tone of the patient based on the calibrated base skin tone image. The system receives a concern image of a portion of the patient's skin, and selects a set of machine learning diagnostic models from a plurality of sets of candidate machine learning diagnostic models based on the base skin tone of the patient, each of the sets of candidate machine learning diagnostic models trained to receive the concern image and output a diagnosis of a condition of the patient.