The International Skin Spectra Archive (ISSA): a multicultural human skin phenotype and colour spectra collection

Introduction

The International Skin Spectra Archive (ISSA) offers a detailed collection of spectral and colorimetric data for human skin, encompassing 15,256 records from 2,113 subjects. This data spans from 2012 to 2024 and originates from eleven datasets curated by international laboratories across eight countries: the UK, Spain, China, Japan, Pakistan, Thailand, Iraq, and Saudi Arabia. Each dataset follows a standardised measurement protocol to maintain data consistency.

In the ISSA dataset, individual records provide extensive details including record number, data origin, subject identification, and skin type—categorised by ethnicity, gender, age, and body location. The dataset also includes detailed information on the measurement instruments used, such as type, specular component inclusion, wavelength range and interval.

Alongside spectral data, each sample also contains CIE colorimetric data, including tristimulus values, xy chromaticity coordinates, CIELAB parameters, etc., based on the CIE 1931 standard colorimetric observer and the CIE standard illuminant D65.

Data Records

The dataset is organised into two primary spreadsheets: a coding scheme and a datasheet. The datasheet arranges data across columns labelled A to BQ:

• A: Unique record identifier
• B: Data origin (linked to an origin table in the coding scheme)
• C: Subject number
• D to G: Ethnicity, gender, age group, and body location (each linked to respective reference tables in the coding scheme)
• H to L: Instrument details including type and spectral measurement specifics
• N to BD: Spectral data from 360 nm to 780 nm
• BF to BQ: CIE colorimetric data

This structured format ensures that users can easily access and interpret data for diverse research applications.

Measurement Protocol

The nationality or ethnicity, gender, age, and body location of the subject are important and depend on the requirements of the specific study that needs the measurement of human skin colour. Thus, this information for each participant was first determined through self-evaluation via a questionnaire. In regions with homogeneous populations (e.g., China, Thailand, Japan), all participants belonged to the same ethnic group. In regions with mixed populations (e.g., the UK), participants were provided with a questionnaire that included options for existing ethnicities, mixed ethnicity, and “other” (self-defined). This information was recorded using consistent coding schemes in the dataset.

To ensure the integrity and consistency of all skin spectral reflectance measurements in this study, several standardised conditions were rigorously maintained. First, it was essential for the skin of all subjects to be clean, unabraded, and free from any cosmetics, lotions, or medical products that could affect the measurement outcomes. Each subject was prepared accordingly prior to data collection to meet this standard. Additionally, the measurement instruments (portable spectrophotometer - SP) were calibrated according to manufacturer guidelines before each session. Lighting conditions for measurements were carefully controlled; all measurements were conducted under diffuse lighting conditions to avoid discrepancies associated with collimated light sources. This was facilitated by lighting systems integrated within the measurement instruments (SP). Such controlled environments guaranteed that the spectral data collected was accurate and consistent across all subjects and datasets.

During the measurement, the portable spectrophotometer was brought to the subject, ensuring the sample area had no blemishes (e.g. hair, freckles, etc.) and had not been subjected to recent pressure (e.g. to promote blood flow). Areas with visible hair (e.g., the chin of male participants with beards) were avoided during measurements to ensure that only clear skin areas were included in the dataset. When using a spectrophotometer, care was taken to ensure the instrument was gently in contact with the area of skin to be measured, to prevent extraneous light from reaching the detector (see Figure 1b). It was also necessary to be careful that no excessive pressure was applied to the skin when contacting the device, which might lead to a change in its colour due to the promotion of blood flow beneath the surface. Measurement parameters mentioned in the last section, such as the geometry of illumination and the inclusion of the specular component, must be checked before measurement. Measurements were taken at 4-10 different body locations depending on the site, with the forehead, cheek, and back of the hand covered by all sites. For consistency, measurements were taken from one randomly selected side (left or right) of each subject. Laterality was generally not recorded, except for Datasets 7 and 8, which included bilateral cheek measurements.

After completing the measurements, we reviewed the data and excluded any measurements with zero reflectance (e.g., due to instrument error) or any participant data with colour difference greater than 15 ∆E_ab^* between positions. All measurement data were then carefully recorded and reported in accordance with the predefined coding schemes. This protocol not only supports the reliability of our measurements but also enhances the comparability of our data across various locations and time periods.

Citing the Database

Any use of the LSDB should cite the following reference: Lu, Yan; Xiao, Kaida; Pointer, Michael; He, Ruili; Zhou, Sicong; Nasseraldin, Ahmed; et al. (2025). The International Skin Spectra Archive (ISSA): a multicultural human skin phenotype and colour spectra collection. figshare. Dataset. https://doi.org/10.6084/m9.figshare.28228571

Contact the Author

Support and Contact Information: For technical support or queries related to the database, please contact Prof. Kaida Xiao (k.xiao1@leeds.ac.uk) and Dr Yan Lu (y.lu3@leeds.ac.uk).