人脸表情识别在智能机器人中的应用研究

摘要

随着智能机器人的快速发展,如何赋予机器人和谐的人机交互能力使其能够感知人类的情感成为当前人机交互研究的热点.针对AAM提取人脸表情特征时表征能力不足和实时性差的问题,提出一种基于BRISK和AAM组合方式提取表情的形状和纹理特征的方法.首先对初始的人脸图像采用Fast-SIC算法拟合出人脸的AAM模型,在获得人脸关键特征点之后用BRISK匹配特征点以增强匹配效率;其次用LGBP对人脸AAM模型的纹理特征进行提取以增强表情特征的表征能力;最后用SVM分类器对提取的表情特征进行分类.实验结果表明,BRISK与AAM组合的特征提取方法可以提高AAM模型的拟合速率,用LGBP提取的纹理特征更具有可分性.在CK+和LFPW人脸库上验证了该算法对面部关键特征点的检测精度和效率,而且与其他算法相比取得了较高的表情识别率,最后在NAO机器人平台上验证了算法的实用性.%With the rapid development of intelligent robots,how to give robots harmonious human-computer interaction makes it able to per-ceive human emotion has been becoming the current hot topics in the study of human-computer interaction.According to the problem of in-sufficient representation of facial expression features and the poor performance of real-time feature extraction using traditional AAM model, we propose a novel method of extracting shape and texture features of facial expression based on the combination of BRISK and AAM.First-ly,the Fast-SIC algorithm is used to fit out the AAM of face for the original face image.In order to enhance the efficiency of feature matc-hing,BRISK algorithm is used to match the acquired key facial feature points,and then LGBP is utilized to extract texture features of AAM in order to strengthen the representation of facial expression features.Finally,to classify the classes of facial expression features using the SVM classifier.The experiments show that the proposed method has improved the efficiency of fitting AAM,and texture features extracted by LGBP have more separability.The detection accuracy and efficiency of key facial feature points are validated respectively on the datasets of CK+and LFPW,and this method has achieved high facial expression recognition rate compared with other algorithms.Finally,its practica-bility is verified on the platform of NAO robot.