Radiance measurement method and system

Abstract

A facial skin radiance score is generated based on skin characteristic metrics extracted from a face image. A plurality of parameters are determined from analysis of a facial skin image, each parameter defining one of a plurality of metrics 103. A trained machine learning model determines a facial skin radiance score in the image using the metrics 105. The skin radiance score is outputted to a display. Image analysis may be used to determine skin characteristics such as colour, luminosity, gloss, texture, pore number, wrinkle depth etc. The machine learning model is trained 105 using a dataset comprising for each of a number of images, a visual radiance score, where the score may be based on a visual assessment of an image by a human assessor, and values of skin characteristic parameters extracted from the image. The trained model determines a radiance score based on the parameters 105. A second score may be determined later, for example to indicate the efficacy of a skincare product.

Images