Challenges in Face Recognition Using Machine Learning Algorithms: Case of Makeup and Occlusions

This paper studies a Face Recognition problem caused by significant variations in ace images, which in practice can be due to different poses, emotion expressions, hairstyles or makeup. Existing Artificial Neural Networks (ANN) have achieved a high recognition accuracy comparable with or even better than human recognition, However in the presence of significant variations the existing ANN methods are still weak. We introduce a new benchmark data set of face images with variable makeup, hairstyles and occlusions, named BookClub artistic makeup data, and then examine the performance of the ANNs under different conditions. In our experiments, the recognition accuracy has decreased when the test images include unseen types of the makeup and occlusions, happened in a real-world scenario. We show that the makeup and other occlusions can be used not only to disguise a person’s identity from the ANN algorithms, but also to spoof a wrong identification.