초록
본 논문은 애니메이터로 하여금 표정공간으로부터 임의의 표정상태 수 개를 선택하도록 하면, 최적의 표정전이경로를 자동적으로 설정하도록 해줌으로써, 얼굴 표정 애니메이션을 실시간적으로 생성하거나 표정 제어가 가능하도록 하기 위한 기법을 기술한다. 표정공간은 약 2500개의 얼굴 표정상태 간의 거리를 구하고, 다차원 스케일링 기법을 사용하여 2차원 평면에 분포시킴으로서 형성된다. 표정공간에서 최적의 표정전이경로를 설정하기 위해서는 임의의 얼굴 표정상태를 기준으로 사분면처럼 4개의 영역으로 나눈다. 그리고 각 영역별로 최단거리에 존재하는 열굴 표정상태를 결정하고, 그 중에서 가장 가까운 얼굴 표정상태를 선택하여 전이시키고, 전이가 끊어진 얼굴 표정상태에서는 두 번째, 세 번째 혹은 네 번째로 가까운 얼굴 표정상태를 선택하여 순서대로 전이시킴으로써 완전한 표정전이경로가 결정된다. 그리고 애니메이터가 표정공간에서 대표적인 수 개의 표정상태만을 선택해주면 시스템은 자동적으로 최적의 표정전이경로를 설정하여 준다. 본 논문은 애니메이터들로 하여금 본 시스템을 사용하여 얼굴 애니메이션을 생성하거나 표정 제어를 수행하도록 하였으며, 그 결과를 평가한다.
This paper presents a facial animation and expression control method that enables the animator to select any facial frames from the facial expression space, whose expression transfer paths the system can setup automatically. Our system creates the facial expression space from approximately 2500 captured facial frames. To create the facial expression space, we get distance between pairs of feature points on the face and visualize the space of expressions in 2D space by using the Multidimensional scaling(MDS). To setup most suitable expression transfer paths, we classify the facial expression space into four field on the basis of any facial expression state. And the system determine the state of expression in the shortest distance from every field, then the system transfer from the state of any expression to the nearest state of expression among thats. To complete setup, our system continue transfer by find second, third, or fourth near state of expression until finish. If the animator selects any key frames from facial expression space, our system setup expression transfer paths automatically. We let animators use the system to create example animations or to control facial expression, and evaluate the system based on the results.