Deep Secure Quantization: On secure biometric hashing against similarity-based attacks.
Signal Processing(2019)
摘要
•This paper proposes a novel deep neural network-based biometric hashing framework.•It offers strong resistance to both conventional ciphertext-only attacks and recent similarity-based attacks.•We analyze why existing biometric hashing models are vulnerable to similarity-based attack.•Our work suggests new definition of non-invertibility in biometric hashing.
更多查看译文
关键词
Random projection,Biometric hashing,Deep learning
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络