Deep Learning Models for Automatic Makeup Detection

Makeup can disguise facial features, which results in degradation the performance of many facial-related analysis systems, including face recognition, landmark characterisation, aesthetic quantification and automated age estimation methods. Thus, makeup is likely to directly affect several real-life applications such as cosmetology virtual cosmetics recommendation security access control, social interaction. In this work, we conduct a comparative study design detection systems leveraging multiple learning schemes from single unconstrained photograph. We have investigated studied efficacy deep models for incorporating use transfer strategy with semi-supervised using labelled unlabelled data. First, during supervised learning, VGG16 convolution neural network, pre-trained on large dataset, fine-tuned Secondly, two unsupervised methods, are self-learning convolutional auto-encoder, trained data then incorporated learning. Comprehensive experiments been conducted 2479 images 446 collected six challenging datasets. The obtained reveal that auto-encoder merged gives best achieving an accuracy 88.33% area under ROC curve 95.15%. promising reflect efficiency combining different strategies by harnessing It would also be advantageous beauty industry develop computational intelligence