A DEEP LEARNING TECHNIQUE FOR FRAUD FACE DETECTION

Abstract

Fraud face detection is a crucial procedure for many face recognition systems. In recent years, state-of-the-art approaches based on convolution neural network (CNNs) show impressive results compared to traditional methods using hand-crafted features. In addition, the increasing trend of embedding the computer vision system on mobile devices requires that the designed algorithms are capable of dealing wit the time-critical constraint. In this paper, we first propose a CNN model, namel hduNet, developed from Google’s MobilenetV2 that provides a flexible trade-of between latency and accuracy, to detect different face spoofing attacks. We then provide an addition dataset of roughly 5000 images capturing the characteristics of Vietnamse people. Combining with LCC_FASD [1] dataset (which is only 1942 read face images, while having 16855 fake face images), the proposed model is carefull fine-tuned to optimize the computational cost as well as the classification accuracy. To validate the model, different experiments have been conducted, demonstrating interesting performance in comparison with other methods.