作者: A. Tanju Erdem
DOI:
关键词: Geography 、 Triangle mesh 、 Rotation (mathematics) 、 Translation (geometry) 、 Motion estimation 、 Computer vision 、 Artificial intelligence 、 Composite image filter 、 Motion (physics) 、 Face (geometry) 、 Facial expression
摘要: A method for tracking the motion of a person's face purpose animating 3-D model same or another person is disclosed. The carries both geometry (shape) and texture (color) characteristics face. shape represented via triangular mesh (geometry mesh), while 2-D composite image (texture image). Both global local are tracked. Global involves rotation translation in 3-D. Local lips, eyebrows, etc., caused by speech facial expressions. positions salient features and/or markers placed on automatically tracked time-sequence images separately calculated using markers. position vectors an action vector that specifies amount actions such as smiling-mouth, raised-eyebrows, etc.