Make-Up Invariant Face Recognition under Uncontrolled Environment

Face Recognition under makeup is a challenging task in the field of biometric. People or criminal identification while having makeup on their faces are important applications. To address this issue, this paper proposes a local descriptor for computing the similarity features for intra-class variation and discriminative features for inter-class similarity. The proposed Min-Max Difference Pattern descriptor is applied over the local patches of the gray scale face image to create the feature vector. A Multi-class Support Vector Machine has been used with 5-fold cross validation for computing recognition accuracy. The proposed descriptor is robust to handle makeup with pose, illumination and expression variation and improved the results. The evaluation of the proposed descriptor has been performed on two standard and challenging datasets-YMU and VMU. The proposed approach outperformed to the descriptor based existing approaches and comparable to some deep learning methods.