Automated Image Analysis in the Evaluation of Facial Pigmentation in Melasma: Clinical Application of Pycnogenol
Journal of Cosmetic Dermatology · 2026-11 · 25 (11)
DOI: 10.1111/jocd.71274 ↗Abstract
ABSTRACT Introduction Melasma is characterized by chronic hyperpigmentation and has a significant psychosocial impact. Traditional clinical assessments are limited due to subjectivity and low sensitivity to detect small changes. Objective To evaluate the progression of facial pigmentation in patients undergoing melasma treatment over 60 days using automated image processing based on artificial intelligence. Methods This prospective clinical trial enrolled 30 individuals with facial melasma into intervention (Pycnogenol 75 mg/day) and control group. Standardized photographs were obtained at baseline (T 0 ), 30 days (T 30 ), and 60 days (T 60 ). Images were processed using a computational pipeline that included automatic renaming, facial segmentation with MediaPipe, extraction of the mean intensity across RGB channels, and application of principal component analysis. Linear mixed‐effects models assessed the effects of group, time, and group‐by‐time interaction. Results PC1 accounted for 99.35% of the variance in gray intensity, indicating a robust marker of pigmentation. A significant effect of time was observed ( p = 0.048), indicating overall changes in pigmentation over the 60‐day period, with greater changes during the first 30 days. Group‐by‐time interaction showed p = 0.056, suggesting a trend toward greater response in the intervention group. This group was also marked by reduced PC1 at T 30 , followed by stabilization at T 60 . Conclusion The automated analysis of facial images proved to be an objective tool for monitoring melasma and detecting subtle changes not captured by subjective clinical methods. Treatment with Pycnogenol showed greater lightening than placebo, particularly during the first month.
EUVIMED