1-HEMT-1-Memristor With Hardware Encryptor for Privacy-Preserving Image Processing

IEEE Electron Device Letters(2022)

引用 3|浏览12
暂无评分
摘要
Efficiency and privacy-preserving are two key requirements in developing computing systems for processing of privacy-related information. Here, we experimentally demonstrate one AlGaN/GaN high-electron-mobility transistor (HEMT) one Au/NbO x /HfO 2 /Pt memristor (1HT1R) structure, which features sub-8ns latency, < 96 fJ programming energy, 10 4 s retention at 85 °C and endurance of >10 6 cycles. We further build a hardware encryptor by combining HEMT-implemented XOR gate and a true random number generator (TRNG) based on variation in Pt/SiO x :Ag/Pt memristor. The proposed TRNG is able to generate 0.48Mb random bitstreams with ultralow energy consumption of 50fJ/bit. A memristors-based homomorphic encryption system is thus constructed based on the 1HT1R and hardware encryptor, capable of performing image segmentation with ultrafast speed, low energy consumption, and privacy-preserving capability.
更多
查看译文
关键词
Memristor,TRNG,encryptor,XOR gate,homomorphic encryption
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要