摘要
视觉交互是自然人机交互的重要组成部分,而人脸特征提取则是视觉交互成功的关键。针对小波变换难以充分描述人脸曲线特征的缺点,为了更好地提取人脸特征,将更符合人类视觉特性的曲波变换用于人脸信息处理,提出了结合曲波变换与Adaboost方法的人脸检测优化方法和基于曲波变换与SVM进行表情分析的新方法,并开展了人脸检测、人脸识别与表情分析的对比实验。实验结果显示,曲波变换在人脸特征提取中具有明显优势,从而为自然人机交互的下一步工作打下了坚实基础。
Vision interaction is one of important aspects of human-computer interaction, and the facial feature extraction is crucial to vision interaction. This paper applies the curvelet transform to the face processing to extract facial feature more effectively. It overcomes the weakness of the wavelet transform which is unable to efficiently extract curve features of face images. An optimized method based on Adaboost and curvelet transform is proposed for face detection. A new approach combining SVM and eurvelet transform is designed for facial expression recognition. Experiments on face detection, face recognition and facial expression recognition are carried out. The results reveal that eurvelet transform has distinct advantages in facial feature extraction, and lays a good foundation for the further work of the natural human-computer interaction.
出处
《中国图象图形学报》
CSCD
北大核心
2010年第9期1309-1317,共9页
Journal of Image and Graphics
基金
国家自然科学基金项目(NSFC-60672071)
浙江省自然科学基金重大项目(D1080807)
关键词
曲波变换
人脸检测
人脸识别
表情分析
人机交互
curvelet transform
face detection
face recognition
facial expression recognition
human-computer interaction
