제목: Blending Face Details: Synthesizing a Face using Multi-Scale Face Models

발표자: 윤승현 교수님

발표자 소속: 동국대학교

날짜: 1/23(월), 2017

시간: 03:00PM - 05:00PM

장소: 301동 417호

호스트: 김명수 교수님

발표내용:

In this talk, we present a novel approach to synthesize a 3D face model from different faces. A 3D face model is parameterized into a 2D parameter space and hierarchically decomposed into a multi-scale face model (MFM). The MFM consists of a base surface, and multi-scale continuous displacement maps (CDMs) describing geometry details from coarse to fine scales. Using multi-scale CDMs, we achieve full correspondences across multiple scales in the parameter space. The salient face regions are marked by 2D feature curves over the parameter space using the intuitive user controls. Then, they drive correspondences across MFMs even with large variations in geometry (e.g. human and animal). Due to the full correspondences across scales and faces, a user can intuitively synthesize a new 3D face model using weighted blending of the MFMs. The method provides semantic control, which means the user can adjust strength or influence of MFMs for a new face model.

발표자 소개:

동국대학교 멀티미디어공학과 부교수(2013. 9 ~ 현재)

동국대학교 멀티미디어공학과 조교수(2007. 9 ~ 2013. 8)

서울대학교 BK21 정보기술사업단 박사후연구원(2007. 3 ~ 2007. 8)

서울대학교 컴퓨터공학부 박사 (2007)

한양대학교 수학과 학사 (2001)