Detection of Electromagnetic Trojans Based on API Cycle Mining

Abstract

Computer electromagnetic radiation Trojan arouses concern gradually due to its particularity in route of transmission. Currently, technologies for preventing electromagnetic radiation are mainly divided into two types of software and hardware. Hardware has a high cost and limited range of application, while most of current software protection mechanism is in theoretical research stage with quite complex principles. This paper makes exploration on prevention methods for displayer electromagnetic Trojan. Trojans need to be realized through changes of pixel on the displayer when working, their function called by their bottom of system would also present characteristics of corresponding sequence. Based on this characteristic, this paper proposes Trojan detection methods based on API sequence cycle mining, and pixel interference is also proposed as a prevention method at the end of this paper. It is indicated by experiments, the method could better detect displayer electromagnetic Trojan programs and has a strong versatility, low cost, and easy to deploy applications.