YouTube Makeup Dataset (YMU)

A face image dataset obtained from YouTube makeup tutorials. The images are captured both before and after the makeup application. The dataset consists of 151 white, female subjects with a minimum of 4 images per subject (2 before-makeup images and 2 after-makeup images). Some subjects have 6 images (3 before-makeup images and 3 after-makeup images). The amount of makeup a subject applies to their face varies from subtle to heavy. The cosmetic alteration is primarily applied to the ocular area; the ocular areas are modified using diverse set of eye makeup products. The skin appearance is also changed due to the application of foundation and lip stick/gloss. In some cases, the hair style changes drastically between the before and after makeup images. The dataset also include expression and pose variations.