OUR MISSION:
OUR MISSION:
This app was developed to provide researchers with an accessible, quick, and efficiently packaged tool to objectively quantify skin color as numerical data from patient images.
Embedded Files
THE PROBLEM:
THE PROBLEM:
Current medical research lacks a robust understanding of skin color as an objective variable, so it can be difficult to quantify its impact as a both a social and physical determinant of health. Currently, the use of a physical colorimeter tool is necessary to collect objective data on skin color. These tools must be used in-person with the patient present, therefore data collection is limited by expensive equipment, timing, transportation, and prospective planning.
Retrospective data collection on skin color may be available in the medical record in the form of digital photos, but it cannot be quantified or objectively extracted from the medical records without expert identification at this time (like a dermatologist or plastic surgeon) but even this is subject to bias and can differ from provider to provider.
OUR SOLUTION:
OUR SOLUTION:
Skin Analyzer was developed to allow us and other researchers to collect objective, retrospective data on Fitzpatrick skin color from the medical record with the aim to further understanding of skin color as a determinant of health without the need for extensive training or expensive equipment. This can be done in a cheaper, faster, objective and reliable manner using the Skin Analyzer.
HOW IT WORKS:
HOW IT WORKS:
This app was created based on the 1976 CIELAB colorspace, which plots color in three axes based on Light/Dark (the L* axis), Red/Green (the a* axis), and Blue/Yellow (the b* axis) and algorithms that correlate CIELAB color space coordinates with human skin pigmentation, erythema, and tanning.
Our app analyzes digital photos of skin and converts an area of pixels in a subjects photo to coordinates in the L*a*b* colorspace. These coordinates are used to calculated the The mean Individual Typology Angle (ITA) which correlates to one of the six Fitzpatrick skin type.
Variables that Can be Collected with Skin Analyzer:
Variables that Can be Collected with Skin Analyzer:
L*a*b* - in the CIELAB color space, colors are assigned coordinates on the Light/Dark (the L* axis), Red/Green (the a* axis), and Blue/Yellow (the b* axis). These coordinates are collected as individual variables.
ITA angle - The individual typology angle equation harnesses colorspace coordinates to calculate skin color "volume", a number that allows for objective classification of skin color for medical research.
Delta E - Using two points on the same subject it calculates the difference between colors using L*a*b* coordinates. A Delta E value greater than one indicates a color difference discernible by the human eye.
Fitzpatrick Skin Type - a scale developed to classify skin coloring and response to ultraviolet light. This scale is widely used to determine eligibility and dose for dermatologic treatments and continues to expand in its application.
RGB - The RGB colorspace is used to define color in digital images and plots color coordinates on a Red, Green, and Blue axis.
x and y coordinates - The x and y coordinates for the center of each region sampled are recorded to allow for the exact pixels used in the Fitzpatrick calculation to be accessed at a later time. This may be used by other softwares, such as Artificial Intelligence (AI), to help identify skin areas significant to the user on an image without the need for another researcher to reidentify these locations.
Disclaimer:
Disclaimer:
This tool is not approved by the FDA as a medical device and should not be used as one. We have only validated its accuracy of obtaining known values and for inter-rater reliability. This preliminary data may be not as accurate as a skin colorimeter and has not yet been validated against one. This tool should not be used for medical advice or treatment and should not be used outside of research purposes or as a substitute for consulting your doctor. By downloading this software you are acknowledging that you have read this disclaimer.
Page updated
Google Sites
Report abuse