Acceso al sitio web del Consejo de Políticas de Infraestructura

   
ARTÍCULO
TITULO

A deep learning approach to face swap detection

Svetlana Volkova    
Alexey Bogdanov    

Resumen

We propose the method for detecting an incident at face authentication when an imposter falsifies the client's face using a digital technique named Face Swap to cheat the system. The method is based on a convolutional neural network to get facial features and classify them. The proposed method can work with faces obtained with low quality and heavy lighting conditions. It is confirmed by experiments on a big test dataset. Experiments show that the accuracy reaches values over 98% for low-quality images and over 99% for high-quality images. Classification results are congruent to the best results shown by the other known methods tested on the same test dataset. The proposed method can be applied to improve the quality of face authentication systems.

PÁGINAS
pp. 16 - 20

 Artículos similares

       
 
Ryota Higashimoto, Soh Yoshida and Mitsuji Muneyasu    
This paper addresses the performance degradation of deep neural networks caused by learning with noisy labels. Recent research on this topic has exploited the memorization effect: networks fit data with clean labels during the early stages of learning an... ver más
Revista: Applied Sciences

 
Giorgio Lazzarinetti, Riccardo Dondi, Sara Manzoni and Italo Zoppis    
Solving combinatorial problems on complex networks represents a primary issue which, on a large scale, requires the use of heuristics and approximate algorithms. Recently, neural methods have been proposed in this context to find feasible solutions for r... ver más
Revista: Algorithms

 
Luis M. de Campos, Juan M. Fernández-Luna, Juan F. Huete, Francisco J. Ribadas-Pena and Néstor Bolaños    
In the context of academic expert finding, this paper investigates and compares the performance of information retrieval (IR) and machine learning (ML) methods, including deep learning, to approach the problem of identifying academic figures who are expe... ver más
Revista: Algorithms

 
Xie Lian, Xiaolong Hu, Liangsheng Shi, Jinhua Shao, Jiang Bian and Yuanlai Cui    
The parameters of the GR4J-CemaNeige coupling model (GR4neige) are typically treated as constants. However, the maximum capacity of the production store (parX1) exhibits time-varying characteristics due to climate variability and vegetation coverage chan... ver más
Revista: Water

 
Yongen Lin, Dagang Wang, Tao Jiang and Aiqing Kang    
Reliable streamflow forecasting is a determining factor for water resource planning and flood control. To better understand the strengths and weaknesses of newly proposed methods in streamflow forecasting and facilitate comparisons of different research ... ver más
Revista: Water