Paper Title
A Deep Convolutional Neural Network with Sharpening Filters for Face Recognition
Abstract
In face recognition problem, human identities can be predicted based on its raw facial features. Traditional
approaches for classification are based on statistical approaches dimension reduction, while the neural networks and support
vector machines on machine learning. However, the performance using conventional features rises dramatically in complex
intra-personal face variations, such as illumination, expression, pose, makeups and occlusions. Therefore, reducing the intrapersonal
variations is an eternal topic in face recognition that is crucial matters. This paper proposes a convolutional neural
network (CNN) embedded with sharpening filters which can improve the problem. This model can classify an input image
into a large number of identity classes, because Sharpening convolutional neural network greatly improve the model
generation capacity by introducing effective Sharpening layer.
Keywords - Biometric Authentication, Deep Learning, Face Rec-ognition, Deep Convolutional Neural Networks,
Sharpening Filter.