Shape analysis for power signal cryptanalysis on secure components

This paper presents an application of pattern recognition techniques in reverse engineering for smart cards. The aim of the study is to design algorithms based on shape classification and to determine instructions executed on a chip as well as processed data sets. Information is extracted from the power consumption in order to recover secret information. Then geometrical features are determined and a syntactic analysis is achieved in order to recover secret algorithms and data. Some examples are given showing how code execution can be reversed on a recent secure component. These examples are essentially focused on instruction recovery but the algorithms also work on data recovery or on a combination of both instruction and data recovery.