Face Recognition Using 3D CNN and Hardmining Loss Function

Abstract

Research in face recognition has achieved new heights after the advent of deep learning, particularly 2D Convolution Neural Network (2D CNN). However, to solve the real-world challenges in face recognition, researchers have started using 3D CNN with video as input. The use of video adds temporal dimension to the input, in addition to the spatial dimension. However, the features extracted by 3D CNN are not as discriminative as those of 2D CNN. In this paper, we propose a framework called $HL_{3}CNN$ framework. The $HL_{3}CNN$ framework uses the Hardmining loss with 3D CNN to increase the discriminative ability of the 3D CNN features. The Hardmining loss samples hard examples in the training dataset to increase the performance of the 3D CNN model. In our experiments, the increased accuracy of the $HL_{3}CNN$ framework to 99.23% on an in-house collected CVBL dataset confirms that the Hardmining loss enhances the discriminative ability of a basic loss function for 3D CNN features.