Small-Window Entropy Based on Fast Multivariate Empirical Mode Decomposition for Characterization of Human Skin Aging

The escalating worldwide demographic trend toward population aging has elevated skin aging to a prominent focus within the realm of health sciences. Traditionally, the assessment of skin aging has primarily relied upon clinical observation and visible alterations on the surface of skin, such as wrinkles, pigmentation spots, and diminished elasticity. However, such approaches often fail to provide adequate objective information and cannot accurately provide dermatologists with early diagnostic information. In recent years, the rapid development of signal decomposition algorithms and high-frequency quantitative ultrasound technology have created new possibilities in the field of skin aging. In this study, being the first in this area, we propose a novel characterization approach for the assessment of human skin aging using 42 MHz high-frequency ultrasound radio frequency (RF) signals. The raw ultrasound RF data are decomposed into 4 intrinsic mode functions (IMFs) using fast multivariate empirical mode decomposition (FMEMD). And 15 different IMFs are obtained by combining 4 individual IMFs. The dermis in all IMFs is determined as the region of interest, and the corresponding parameter values are calculated based on the small-window entropy, respectively. All of the above are performed on the customized dataset containing high-frequency ultrasound RF signals from the facial skin of 60 female individuals aged from 23.8 to 56.8 years old. The results suggest that the decomposed small-window entropy based on IMF<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1,2,3</inf> has great feasibility in assessing human skin aging.