Cryptology ePrint Archive: Report 2018/177

On the Use of Independent Component Analysis to Denoise Side-Channel Measurements

Houssem Maghrebi and Emmanuel Prouff

Abstract: Independent Component Analysis (ICA) is a powerful technique for blind source separation. It has been successfully applied to signal processing problems, such as feature extraction and noise reduction, in many different areas including medical signal processing and telecommunication. In this work, we propose a framework to apply ICA to denoise side-channel measurements and hence to reduce the complexity of key recovery attacks. Based on several case studies, we afterwards demonstrate the overwhelming advantages of ICA with respect to the commonly used preprocessing techniques such as the singular spectrum analysis. Mainly, we target a software masked implementation of an AES and a hardware unprotected one. Our results show a significant Signal-to-Noise Ratio (SNR) gain which translates into a gain in the number of traces needed for a successful side-channel attack. This states the ICA as an important new tool for the security assessment of cryptographic implementations.

Category / Keywords: independent component analysis, side-channel analysis, preprocessing, noise filtering, correlation power analysis, Boolean masking scheme

Original Publication (with minor differences): COSADE 2018

Date: received 12 Feb 2018

Contact author: houssem mag at gmail com

Available format(s): PDF | BibTeX Citation

Version: 20180214:125008 (All versions of this report)

Short URL: ia.cr/2018/177

Discussion forum: Show discussion | Start new discussion


[ Cryptology ePrint archive ]