基于RBF神经网络的人形机器人在线面部表情模仿

doi:10.13973/j.cnki.robot.2016.0225

摘要
图/表
参考文献
相关文章 (1)

全文: PDF (647 KB) HTML ( ) 输出: BibTeX \| EndNote (RIS)
引用本文:
黄忠, 任福继, 胡敏. 基于RBF神经网络的人形机器人在线面部表情模仿[J]. 机器人, 2016, 38(2): 225-232.DOI: 10.13973/j.cnki.robot.2016.0225. HUANG Zhong, REN Fuji, HU Min. Online Facial Expression Imitation for Humanoid Robot Based on RBF Neural Network. ROBOT, 2016, 38(2): 225-232. DOI: 10.13973/j.cnki.robot.2016.0225.

摘要

针对有限数目电机以及手工设置面部表情控制参数的局限，结合基于 Kinect 的主动外观模型，提出一种基于径向基函数神经网络的人形机器人在线面部表情模仿算法．在离线面部表情学习阶段，基于径向基函数网络建立前向机械模型以反映电机控制值与表情形变特征的映射关系，并进一步构建逆向预测模型以规整电机连续运动的平滑度；在在线面部表情模仿阶段，基于前向机械模型和逆向预测模型寻找最优电机值以实现机器人与表演者形变偏差的最小化，并引入权重因子调节表情模仿的瞬时相似度和电机连续运动的平滑度．最后，从均值统计和预测偏差角度验证两模型的合理性和泛化能力，并进一步讨论了权重因子对时空相似性和平滑度的影响．实验结果表明：前向机械模型形变预测偏差不超过 1%，逆向预测模型电机控制偏差不超过 1.5%．与 Jaeckel、Trovato、Magtanong 三种方法相比，本文算法在单帧表情模仿相似度以及多帧表情动作平滑度方面均具较好优势．

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	黄忠
	任福继
	胡敏

关键词 ：人形机器人, 径向基函数神经网络, 在线面部表情模仿, 面部表情相似度, 动作平滑度

Abstract：

To overcome the limitations in manually setting the control parameters of facial expressions with motors of limited number, an online facial expression imitation algorithm is proposed for humanoid robot based on RBF (radial basis function) neural network by combining the Kinect based AAM (active appearance model). In the offline facial expression learning phase, a forward mechanic model is modeled based on RBF networks to reflect the mapping relationship between the motor control values and the facial deformation characteristics, and an inverse prediction model is further developed for wrapping the smoothness of continuous motor movements. In the online facial expression imitating phase, optimal motor values are solved to minimize deformation deviations between the robot and the performer, based on the forward mechanic model and the inverse prediction model; moreover, a weighting factor is introduced to adjust the instantaneous similarity of expression imitation and the smoothness of motor's continuous motion. Finally, the rationality and generalization ability of the two models are validated from the perspective of mean statistics and prediction deviations, and the influence of weighting factor on space-time similarity and smoothness is further discussed. The experimental results indicate that the deformation deviations of the forward mechanic model are less than 1%, and the motor control deviations of the inverse prediction model are less than 1.5%. Compared with the three methods of Jaeckel, Trovato and Magtanong, the proposed algorithm has advantages in the similarity of single-frame expression imitation and the smoothness of multi-frame expression motion.

Key words： humanoid robot RBF (radial basis function) neural network online facial expression imitation facial expression similarity motion smoothness

收稿日期: 2015-07-29 录用日期: 2016-01-14

TP249

基金资助:

国家自然科学基金（61432004）；国家自然科学基金青年科学基金（61300119）；情感计算与先进智能机器安徽省重点实验室开放课题（ACMIM150106）；安徽省教育厅自然科学研究项目（AQKJ2015B013）

通讯作者: 黄忠，huangzhong3315@163.com E-mail: huangzhong3315@163.com

作者简介: 黄忠（1981-），男，博士生，讲师.研究领域：人机情感交互，机器人控制.
任福继（1959-），男，博士，教授.研究领域：自然语言处理，人形机器人，人机情感交互.
胡敏（1967-），女，博士，教授.研究领域：计算机视觉，情感计算，人脸表情识别.

链接本文:

https://robot.sia.cn/CN/10.13973/j.cnki.robot.2016.0225 或 https://robot.sia.cn/CN/Y2016/V38/I2/225

[1] Manohar V, Crandall J W. Programming robots to express emotions: Interaction paradigms, communication modalities, and context[J]. IEEE Transactions on Human-Machine Systems, 2014, 44(3): 362-373.

[2] Tadesse Y. Actuation technologies for humanoid robots with facial expressions (HRwFE)[J]. Transaction on Control and Mechanical Systems, 2013, 2(7): 337-349.

[3] 任福继.机器人云与机器人学校[J].科技导报, 2012, 30(9):73-79.Ren F J. Robotic cloud and robotic school[J]. Science & Technology Review, 2012, 30(9): 73-79.

[4] Jaeckel P, Campbell N, Melhuish C. Facial behavior mapping -- From video footage to a robot head[J]. Robotics and Autonomous Systems, 2008, 56(12): 1042-1049.

[5] 刘遥峰, 王志良.基于情感交互的仿人头部机器人[J].机器人, 2009, 31(6):493-500.Liu Y F, Wang Z L. Humanoid head robot based on emotional interaction[J]. Robot, 2009, 31(6): 493-500.

[6] Shayganfar M, Rich C, Sidner C. A design methodology for expressing emotion on robot faces[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA: IEEE, 2012: 4577-4583.

[7] Ahn H S, Lee D W, Choi D, et al. Development of an incarnate announcing robot system using emotional interaction with humans[J]. International Journal of Humanoid Robotics, 2013, 10(2): 1-24.

[8] Wilbers F, Ishi C, Ishiguro H. A blendshape model for mapping facial motions to an android[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA: IEEE, 2007: 542-547.

[9] Magtanong E, Yamaguchi A, Takemura K, et al. Inverse kinematics solver for android faces with elastic skin[C]//Latest Advances in Robot Kinematics. Berlin, Germany: Springer-Verlag, 2012: 181-188.

[10] Trovato G, Zecca M, Kishi T, et al. Generation of humanoid robot's facial expressions for context-aware communication[J]. International Journal of Humanoid Robotics, 2013, 10(1): doi, 1350013-1-22.

[11] 王巍, 王志良, 郑思仪, 等.一类具有相同结构的表情机器人共同注意方法[J].机器人, 2012, 34(3):265-274.Wang W, Wang Z L, Zheng S Y, et al. Joint attention for a kind of facial expression robots with the same structure[J]. Robot, 2012, 34(3): 265-274.

[12] Microsoft. Develop for Kinect[EB/OL]. [2013-12-05]. http:// www.microsoft.com/zh-cn/kinectforwindows/.

[13] 黄忠, 胡敏, 王晓华.一种基于几何特征的表情相似性度量方法[J].模式识别与人工智能, 2015, 28(5):443-451.Huang Z, Hu M, Wang X H. A similarity measurement method of facial expression based on geometric feature[J]. Pattern Recognition and Artificial Intelligence, 2015, 28(5): 443-451.

[14] 樊劲辉, 贾松敏, 李秀智.基于 RBF 神经网络的全向智能轮椅自适应控制[J].华中科技大学学报:自然科学版, 2014, 42(2):111-115.Fan J H, Jia S M, Li X Z. Adaptive control for omni-directional intelligent wheelchairs based on RBF neural network[J]. Journal of Huazhong University of Science and Technology: Natural Science Edition, 2014, 42(2): 111-115.

[15] 樊兆峰, 马小平, 邵晓根.非线性系统 RBF 神经网络多步预测控制[J].控制与决策, 2014, 29(7):1274-1278.Fan Z F, Ma X P, Shao X G. RBF neural network multi-step predictive control for nonlinear systems[J]. Control and Decision, 2014, 29(7): 1274-1278.

[16] Montoliu R, Pla F. Generalized least squares-based parametric motion estimation[J]. Computer Vision and Image Understanding, 2009, 113(7): 790-801.

[17] Tadesse Y, Priya S. Graphical facial expression analysis and design method: An approach to determine humanoid skin deformation[J]. Journal of Mechanisms and Robotics, 2012, 4(2): 1-16.

[18] 朱特浩, 赵群飞, 夏泽洋.利用 Kinect 的人体动作视觉感知算法[J].机器人, 2014, 36(6):647-653. Zhu T H, Zhao Q F, Xia Z Y. A visual perception algorithm for human motion by a Kinect[J]. Robot, 2014, 36(6): 647-653.