Quantification of erythema using digital camera and computer‐based colour image analysis: a multicentre study

Abstract

Background/purpose Colour measurements obtained from digitized images have been proposed as a simple and cost‐effective way to evaluate skin colour and the activity of treatments. The main disadvantage of the method is the fact that it is highly dependent on ambient light: even if an accurate control of subjects' illumination is provided, readings remain not comparable among different laboratories. The purpose of this study was to develop a highly reproducible system for computerized colour image analysis of skin erythema, making it possible to compare readings from different environmental light conditions. Patients and Methods Three hundred and forty‐eight Caucasian adult healthy subjects (age range: 18–60 years) of both sexes (14% males, 86% females), were enrolled in the study by 49 dermatologists distributed all over Italy. They were recruited among patients who required aesthetic treatments involving skin erythema, like chemical peeling and laser epilation. Once the treatment was administered, clinical evaluations and pictures were taken at the level of treated areas. Visual assessment of erythema was done on the basis of conventional clinical grades (0 = absent; 1 = slight; 2 = moderate; 3 = intense). The clinicians participating in the study were asked to put a standard colour marker (red, green and blue coloured self‐adhesive ring) in the photographed skin area. The difference between r, g, b values of photographed colour markers on the skin of single patients participating in the study and the r, g, b values obtained photographing the colour marker in fixed illumination conditions was used to adjust skin colour measurements. Then erythema index (E.I) on digitized images was calculated subtracting red value to green one by averaging procedure of different pixels. Results Erythema index. average value among the groups divided according to the conventional clinical score increased progressively from score 0–2, while it decreased from score 2 to score 3. The differences in E.I. mean values among the score groups (0 vs. 1, 1 vs. 2, 2 vs. 3) were statistically significant (P < 0.05). Conclusion We developed a method for the measurement of skin erythema using digital camera, normalized r, g, b colour co‐ordinate system and computerized calculation of E.I. Clinical usefulness of our method for absent, slight and moderate erythema, was demonstrated. For intense erythema lesions we did not find a correspondence between clinical and computerized evaluation, probably due to other factors involved in skin inflammation (e.g. oedema).