Overview
Explore a 15-minute IEEE conference talk on a novel physical impersonating attack called ImU, designed to exploit face recognition systems through natural style changes. Learn about the research conducted by a team from Purdue University, Nanjing University, Rutgers University, and Clarifai Inc. Discover how this attack method manipulates facial features to deceive recognition algorithms while maintaining a natural appearance. Gain insights into the potential vulnerabilities of current face recognition technologies and the implications for security systems relying on facial authentication.
Syllabus
ImU: Physical Impersonating Attack for Face Recognition System with Natural Style Changes
Taught by
IEEE Symposium on Security and Privacy