An Improved Principal Component Analysis for Side-Channel Attacks
Haoming Bai, Hongling Gao, Shan Yu, Teng Zhai, Ziyang Ma, Qingshuai Guo · 2022 IEEE 10th International Conference on Information, Communication and Networks (ICICN) · 2022
Side-channel attacks (SCAs) often need to collect a lot of power consumption data, and it takes a great deal of time to crack the secret key through these data. Ordinary principal component analysis (PCA) compression techniques consume a lot of time when processing power traces containing a large number of sample points. In this paper, two improved algorithms, absolute value integration PCA (APCA) and maximum value extraction PCA (MPCA), are proposed to optimize PCA. Both can compress the original power traces to about 10%. Compared with PCA, their processing speed is significantly improved. The more sample points, the more obvious the improvement. And through experiments, it is proved that APCA has more advantages than MPCA in the convergence of partial guessing entropy (PGE).