DLDDO: Deep Learning to Detect Dummy Operations

JongHyeok Lee; Dong-Guk Han

Paper 2020/939

DLDDO: Deep Learning to Detect Dummy Operations

JongHyeok Lee and Dong-Guk Han

Abstract

Recently, research on deep learning based side-channel analysis (DLSCA) has received a lot of attention. Deep learning-based profiling methods similar to template attacks as well as non-profiling-based methods similar to differential power analysis have been proposed. DLSCA methods have been proposed for targets to which masking schemes or jitter-based hiding schemes are applied. However, most of them are methods for finding the secret key, except for methods for preprocessing, and there are no studies on the target to which the dummy-based hiding schemes or shuffling schemes are applied. In this paper, we propose a DLSCA for detecting dummy operations. In the previous study, dummy operations were detected using the method called BCDC, but there is a disadvantage in that it is impossible to detect dummy operations for commercial devices such as an IC card. We consider the detection of dummy operations as a multi-label classification problem and propose a deep learning method based on CNN to solve it. As a result, it is possible to successfully perform detection of dummy operations on an IC card, which was not possible in the previous study.

Metadata

Available format(s): PDF
Category: Applications
Publication info: Published elsewhere. The 21st World Conference on Information Security Applications
DOI: 10.1007/978-3-030-65299-9_6
Keywords: Hiding Countermeasure Deep Learning Multi-label classification IC Card Dummy Operation
Contact author(s): n_seeu @ kookmin ac kr
christa @ kookmin ac kr
History: 2020-12-23: last of 2 revisions; 2020-07-31: received; See all versions
Short URL: https://ia.cr/2020/939
License: CC BY

BibTeX

@misc{cryptoeprint:2020/939,
      author = {JongHyeok Lee and Dong-Guk Han},
      title = {{DLDDO}: Deep Learning to Detect Dummy Operations},
      howpublished = {Cryptology {ePrint} Archive, Paper 2020/939},
      year = {2020},
      doi = {10.1007/978-3-030-65299-9_6},
      url = {https://eprint.iacr.org/2020/939}
}