Trojan Scanner: Detecting Hardware Trojans with Rapid Imaging Combined with Image Processing and Machine Learning

Tuesday, October 30, 2018: 1:00 PM
Exhibit Halls A/B (Phoenix Convention Center)
Mr. Nidish Vashistha , University of Florida, Gainesville, FL
Hangwei Lu , University of Florida, Gainesville, FL
Qihang Shi , University of Florida, Gainesville, FL
Mr. M. Tanjidur Rahman , University of Florida, Gainesville, FL
Dr. Haoting Shen, PhD , University of Florida, Gainesville, FL
Damon L. Woodard , University of Florida, Gainesville, FL
Dr. Navid Asadi , University of Florida, Gainesville, FL
Prof. Mark Tehranipoor, PhD , University of Florida, Gainesville, FL

Summary:

Outsourcing of semiconductor manufacturing has exposed the integrated circuits (ICs) to a new threat called hardware Trojans. A hardware Trojan is an intentional modification of an IC circuitry to leak valuable data or disable an entire chip or components of it. In literature, many electrical tests techniques have been proposed to detect hardware Trojans but these techniques cannot detect Trojans with a high confidence level. Also, the field of Trojan detection using image processing has not been much explored in detail. Therefore, in this paper, we introduce an intelligent technique using customized advanced image processing to detect the smallest change in the active region of the chip using golden layout (GDSII). This technique is efficient enough to detect biases due to process variation and defects by using machine learning.