Eyelash Image Characterization Using U-Net for Precise Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for evaluating eyelash characteristics, such as length and density, are cumbersome and require manual operations, limiting user interaction and accuracy, and existing computer vision techniques fail to effectively identify and characterize eyelashes due to their small size and complexity.
Innovation Solution
Employing machine vision techniques, specifically U-Net neural networks, to analyze close-up images of eyelashes, optimizing image acquisition and processing to quantify parameters like length, density, and curvature, and recommending suitable cosmetic products based on these characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual evaluation methods with rulers are used to measure eyelash parameters, then measurement can be performed, but the process is cumbersome, requires multiple positioning and reading operations, and severely limits user interaction
Solution Approach 1:
The patent replaces manual mechanical measurement tools (rulers and positioning devices) with an automated digital imaging system that uses a camera to capture eyelash images and software to automatically measure parameters such as length, density, and curvature. This substitution eliminates the need for manual positioning and reading operations while maintaining measurement precision.
Solution Approach 2:
The system enables self-service measurement by allowing users to simply position their eye in front of the camera, after which the automated software performs all measurement operations without requiring user intervention for positioning or reading. The system serves itself by automatically capturing images, processing them, and generating measurement results.
2Extent of automation
If existing computer vision techniques are used to detect eyelashes, then image processing can be performed, but the techniques fail to effectively identify and characterize eyelashes due to their small size and complexity
Solution Approach 1:
The patent applies local quality by optimizing the imaging system specifically for eyelash detection, including using appropriate magnification, lighting conditions, and image processing algorithms tailored to the small size and specific characteristics of eyelashes. This localized optimization enables accurate identification and characterization of eyelashes that general computer vision techniques miss.
Solution Approach 2:
The system changes key parameters such as image resolution, magnification level, and processing algorithms to be specifically suited for detecting small eyelash structures. By adjusting these parameters, the system achieves both high automation and precise measurement of eyelash characteristics.
3Adaptability or versatility
If multiple products are tried to find the appropriate mascara for desired effects, then product suitability can be determined, but the process is complex and may discourage consumers
Solution Approach 1:
The system performs preliminary characterization of the user's eyelash parameters (length, density, curvature) before product recommendation. This preliminary information is then used to directly recommend suitable mascara products, eliminating the need for consumers to try multiple products to find the right match.
Solution Approach 2:
The system provides feedback to consumers about their eyelash characteristics and recommends specific products tailored to their needs. This feedback loop enables informed product selection without requiring multiple trial products, saving consumers time while maintaining accurate product-adaptability matching.
Data Source
Figure 1
AI summary
The present invention relates to a method for characterizing eyelashes or eyebrows which is implemented by computer and comprises the following steps, which are directed toward: - receiving data corresponding to a set of pixels in a close-up image of a body area comprising a plurality of eyelashes, preferably the entirety of a row of eyelashes or of an eyebrow, to be characterized; - applying at least one step of computer vision so as to obtain, on the basis of the image data received, a numerical evaluation of at least one characteristic numerical parameter from among, notably, the number of fibers, an average length of the fibers, an average thickness of the fibers, and a curvature of the fibers.