Effects of Partially/Masked Face on the Personality Identifying: Self-Multimodal Fusion Consideration

Abstract

Due to the essential using of mask face in various fields and critical situations, there is urgent need to develop techniques for identifying personality to be more effective in presence the mask face. Also, Muslim women who wear Higab, they suffer from the traditional identifying processes. Facial recognition techniques have become ineffective with face mask. In this research, the technique of identifying the personality through the face with the mask was proposed by making a fusion between both the original face and part of the face, using Score fusion and features fusion. The research presented an extensive study of the effect of the mask on the difficulty of recognizing the face and changing the facial features. The research used more than one tool to extract the features of a face image with a mask, and a group of classification tools to identify the personality more accurately. More than one extraction tool has been combined to increase the classification accuracy. The different feature extracting of voice and masked face images are used with employing the various classifiers such as the Support Vector Machine (SVM), K-Nearest Neighbours (KNN), Gaussian Mixture Model (GMM) and Artificial Neural Network (ANN) tools. The simulation experiments results prove the superiority of score based fusion.