Bilkent University
Department of Computer Engineering
CS690 SEMINAR

 

Analyzing and Transferring Personality in Real-Life Images

 

Sinan Sonlu
PhD Student
(Supervisor: Prof. Dr. Uğur Güdükbay)
Computer Engineering Department
Bilkent University

Automatic analysis of real-life images and videos for understanding the apparent personality of individuals is an active area of research with applications in affective computing, human-machine interfaces, gaming, security, and marketing. Previous research suggests the apparent personality of an individual is related to his/her posture, gesture, facial expression, vocal features, and choice of words. Recent advances in pose transfer enable automatic modifications to the body pose of source images. Using the OCEAN personality model as a framework, we focus on analyzing the apparent personality of source images for transferring the acquired features onto target images. We utilize Laban Movement Analysis to define high-level features for the posture. The resulting output will be a substitute for the target image that expresses the traits of the source image.

 

DATE: 21 October 2020, Wednesday @ 13:15