Keratinous Surface Characteristic Detection Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for diagnosing and treating keratinous surfaces, such as hair, face challenges in obtaining reliable and objective data due to the specific texture and environment of hair and scalp, particularly for thin, white, and blond hairs, which are difficult to identify and count accurately, and existing devices like spectrophotometers average values over large areas, including skin and hair, leading to inaccurate measurements.
Innovation Solution
A method using machine learning models to process high-resolution images of the keratinous surface, applying image segmentation and returning numerical values for characteristics like hair tone, density, and other properties, allowing for quantitative determination and personalized product recommendations or compositions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual or visual assessment methods are used by experts, then the process is simple and accessible, but the reliability and objectivity of the diagnosis is poor
Solution Approach 1:
The patent replaces manual visual assessment by experts with an automated image processing system using machine learning models. The system captures images of the keratinous surface and uses trained models to objectively determine characteristics such as color, density, and texture, eliminating subjectivity while maintaining accessibility through standard imaging devices.
Solution Approach 2:
The system enables automated self-diagnosis without requiring expert intervention. The machine learning models process images and automatically generate diagnostic results and product recommendations, allowing the system to serve itself in making accurate assessments that previously required trained professionals.
2Measurement precision
If spectrophotometers are used to measure hair color, then objective measurement is achieved, but the measurement accuracy deteriorates due to averaging over large areas including skin and hair
Solution Approach 1:
The patent applies image segmentation techniques to separate hair from skin in captured images. The machine learning models identify and segment individual hair strands, allowing precise measurement of hair characteristics without contamination from skin areas. This enables accurate detection of light-colored and thin hairs that were previously difficult to distinguish.
Solution Approach 2:
The system transitions from spectral measurement in one dimension to spatial image analysis in two dimensions. By capturing images and analyzing spatial distribution of hair characteristics, the system can distinguish individual hair strands from skin background, providing precise measurements without the averaging problem of spectrophotometers.
3Measurement precision
If high-resolution images are processed with machine learning models, then diagnostic precision is improved, but processing complexity increases
Solution Approach 1:
The patent uses pre-trained machine learning models that have been trained in advance on large datasets. This preliminary training allows the models to quickly and accurately process new images without requiring complex real-time computation. The pre-trained models can be deployed on standard devices, reducing processing complexity while maintaining high precision.
Data Source
AI summary
The present application is directed to a method and system for determining at least one physical and/or chemical characteristic of a keratinous surface of a user, the method comprising the steps of: —receiving data corresponding to at least one image of the keratinous surface, —processing the image by applying at least one machine learning model to said image, —returning at least one numerical value corresponding to a grade of the characteristic of the keratinous surface to be determined.