Local Adaptive Gradient Variance Attack for Deep Fake Fingerprint Detection

作     者：Chengsheng Yuan Baojie Cui Zhili Zhou Xinting Li Qingming Jonathan Wu 

作者机构：Engineering Research Center of Digital ForensicsMinistry of EducationNanjing University of Information Science and TechnologyNanjing210044China School of Computer ScienceNanjing University of Information Science and TechnologyNanjing210044China Institute of Artificial Intelligence and BlockchainGuangzhou UniversityGuangzhou510006China School of International RelationsNational University of Defense TechnologyNanjing210039China Department of Electrical and Computer EngineeringUniversity of WindsorWindsorN9B 3P4Canada 

出 版 物：《Computers, Materials & Continua》 (计算机、材料和连续体（英文）)

年 卷 期：2024年第78卷第1期

页      面：899-914页

核心收录：

学科分类：08[工学] 080203[工学-机械设计及理论] 0802[工学-机械工程] 

基　　金：supported by the National Natural Science Foundation of China under Grant(62102189,62122032,61972205) the National Social Sciences Foundation of China under Grant 2022-SKJJ-C-082 the Natural Science Foundation of Jiangsu Province under Grant BK20200807 NUDT Scientific Research Program under Grant(JS21-4,ZK21-43) Guangdong Natural Science Funds for Distinguished Young Scholar under Grant 2023B1515020041 

主　　题：FLD adversarial attacks adversarial examples gradient optimization transferability 

摘      要：In recent years,deep learning has been the mainstream technology for fingerprint liveness detection(FLD)tasks because of its remarkable ***,recent studies have shown that these deep fake fingerprint detection(DFFD)models are not resistant to attacks by adversarial examples,which are generated by the introduction of subtle perturbations in the fingerprint image,allowing the model to make fake *** of the existing adversarial example generation methods are based on gradient optimization,which is easy to fall into local optimal,resulting in poor transferability of adversarial *** addition,the perturbation added to the blank area of the fingerprint image is easily perceived by the human eye,leading to poor visual *** response to the above challenges,this paper proposes a novel adversarial attack method based on local adaptive gradient variance for *** ridge texture area within the fingerprint image has been identified and designated as the region for perturbation ***,the images are fed into the targeted white-box model,and the gradient direction is optimized to compute gradient ***,an adaptive parameter search method is proposed using stochastic gradient ascent to explore the parameter values during adversarial example generation,aiming to maximize adversarial attack *** results on two publicly available fingerprint datasets show that ourmethod achieves higher attack transferability and robustness than existing methods,and the perturbation is harder to perceive.