New Hyperspectral Imaging Database Enhances Human Skin Research

Scientific facial skin hyperspectral imaging measurements ©fizkes - stock.adobe.com

A team from the Physics Center of Minho and Porto Universities (CF-UM-UP) at the University of Minho (Portugal), led by Andreia E. Gomes, Sérgio M. C. Nascimento, and João M. M. Linhares, has created a hyperspectral imaging database to capture the spectral characteristics of human facial skin. This comprehensive dataset is intended to support a variety of applications, from medical diagnostics to digital rendering, by providing precise, non-invasive measurements of skin reflectance. The research, published in Applied Spectroscopy, offers new insights into the spectral variations in human skin across different tones and regions of the face (1).

Background and Importance of Skin Spectral Data
Human skin color plays a crucial role in perceptions of beauty, health, and emotional status, and varies significantly based on factors like melanin and hemoglobin concentration. Skin color also plays a role in the use of medical devices relying on non-invasive spectroscopic measurements through the skin. Until now, many skin color studies have relied on local point-contact spectrometers, which can capture data from small areas of the skin, but often fail to provide the spatial context necessary for comprehensive analysis (1). Hyperspectral imaging, on the other hand, offers a non-invasive method to record both spectral and spatial data, preserving detailed information about skin color and its variations across the face (1,2).

However, existing databases of skin reflectance data tend to be limited in scope, often focusing only on light skin tones or failing to control for factors like lighting and facial movement. The new database created by Gomes, Nascimento, and Linhares aims to fill these gaps, offering a more diverse, controlled, and comprehensive dataset.

Details of the Hyperspectral Imaging Database
The research team measured the spectral reflectance of 29 participants with different skin tones, ensuring careful control over lighting conditions and facial movements. Participants’ faces were cleaned before measurements, and those with facial hair or visible skin anomalies were excluded from the study. The color of each participant’s skin was classified using both the von Luschan Chromatic Scale, which contains 36 color samples, and the Fitzpatrick Scale, which categorizes skin types based on their reaction to sunlight.

The von Luschan Chromatic Scale and the Fitzpatrick Scale are both tools used to classify human skin color, but they serve different purposes. The von Luschan chromatic scale (VLCS), developed in the late 19th century, consists of 36 colored tiles used to visually compare skin tones, mostly in anthropological and medical research. However, it has fallen out of favor due to its subjective nature and difficulty in consistent use (3). The Fitzpatrick Scale, introduced in 1975, classifies skin types based on their response to UV exposure and tendency to tan or burn. It includes six categories, from very fair skin (Type I) to very dark skin (Type VI), and is widely used in dermatology to assess sunburn risk and plan appropriate treatments (4).

Participants were divided into two groups based on their skin tones. Group 1 included lighter skin tones (von Luschan scores 1-15 and Fitzpatrick phototypes I-III), while Group 2 included darker skin tones (von Luschan scores 16-36 and Fitzpatrick phototypes IV-VI). This approach allowed the researchers to analyze both intergroup and intragroup variations in skin reflectance.

The hyperspectral imaging system used for the study captured data in 10-nanometer steps across the visible light spectrum, from 400 to 720 nanometers. This detailed resolution enabled the team to analyze subtle variations in skin reflectance often missed by conventional methods (1,2).

Learn More: Hyperspectral Imaging

Key Findings and Implications
One of the study’s key findings was that skin reflectance varies not only between individuals with different skin tones, but also across different regions of the face. These local variations in spectral reflectance were significant, particularly when compared to the average values obtained by point-contact spectrometers. This highlights the limitations of traditional measurement methods, and underscores the importance of using hyperspectral imaging for more accurate skin color analysis (1).

The database is expected to be a valuable resource for a range of fields. In psychophysics, it can help improve understanding of how humans perceive skin color and how this perception is influenced by lighting and other factors. In digital media, the data can be used to create more realistic skin models for animations and video games. Additionally, the database has potential applications in medical diagnostics, where subtle changes in skin color can indicate underlying health conditions (1).

Future Applications and Conclusion
The hyperspectral imaging database created by Gomes, Nascimento, and Linhares is a significant advancement in the study of human skin. By providing detailed spectral data for a diverse range of skin tones, the database has the potential to improve applications in fields as varied as psychophysics, dermatology, cosmetic development, non-invasive medical measurements, and artificial intelligence (1).

As hyperspectral imaging technology continues to advance, this database may also support the development of new diagnostic tools and treatments for skin disorders, as well as innovations in skincare products and makeup. By capturing the complex interplay between light and human skin, the researchers have opened up new possibilities for understanding and modeling one of the most important aspects of human appearance (1,2).

References

(1) Gomes, A. E.; Nascimento, S. M. C.; Linhares, J. M. M. Hyperspectral Imaging Database of Human Facial Skin. Appl. Spectrosc. 2024, 0 (0), September 24, 2024. DOI: 10.1177/00037028241279323

(2) Stamatas, G. N.; Balas, C. J.; Kollias, N., Hyperspectral Image Acquisition and Analysis of Skin. In Spectral Imaging: Instrumentation, Applications, and Analysis II, SPIE, 2003, 4959, 77–82). DOI: 10.1117/12.479491

(3) Von Luschan, F. Beiträge zur Völkerkunde der Deutschen Schutzgebieten. Berlin: Deutsche Buchgemeinschaft, 1897.

(4) Fitzpatrick T. B. Soleil et peau [Sun and skin]. Journal de Médecine Esthétique 1975, 2, 33–34.