摘要
设计并研制了具有面部表情和对人类表情识别机能的仿人头像机器人系统。该机器人具有弹性面部皮肤机构,能够实现从CCD图像采集装置获得人类的自然、开心、生气、悲伤、惊讶、厌恶等表情图像,然后进行表情识别,并根据识别结果,对机器人面部皮肤和头部器官进行运动控制再现表情;提出了用于实现机器人表情再现控制的表情特征点偏移量及偏移系数概念和计算方法,解决了图像不同比例、人与机器人面部差异情况下机器人表情再现控制问题;实验结果表明,该机器人能够逼真地再现出所识别的人类6种基本表情。
The human head imitation portrait robotic system that possesses facial expression and recognition enginery on human expressions was designed and developed. This robot has a mechanism of elastic facial skin and could realize the natural, joyous, getring angry, sorrowful, open-eyed, disgusting etc. expression pictures of mankind obtained from the CCD pictures collection device. And then carried through the recognitions on facial expressions, and in accordance with the result of recognition to carry out motion control and reproduce the expression. The computation method on offset of expression character point and concept of excursion coefficient used for realizing the control of robotic expression reproduction were brought forward. The problems of expression reproduction control of robot under the situations of different scale of portrait and the facial diversity between human and robot. The experimental result indicated that this robot could reproduce realistically the being recognized six kinds of basic expressions of mankind.
出处
《机械设计》
CSCD
北大核心
2007年第7期20-24,共5页
Journal of Machine Design
基金
哈尔滨工业大学跨学科交叉基金资助项目(HIT.DM.2002.06)
关键词
仿人头像机器人
面部表情
视觉
表情识别
human head imitation portrait robot
facial expression
vision
expression recognition