摘要
本文提出了基于三维不变矩的人脸表情识别系统研究框架。在对三维人脸表情模型进行预处理及归一化的基础上建立具有平移比例和旋转不变性的三维人脸表情模型的矩不变量特征表示。针对人脸表情分类应用背景采用基于支持向量机的智能学习技术研究实现利用多维度3D-Zernike描述子特征的人脸分析识别系统的建模,基于D-ABC算法的分析识别模型优化,提高模型可靠性和系统的鲁棒性,进而利用该系统进行三维人脸表情数据的分类识别。
This paper presents the research framework of 3D facial expression recognition system based on moment invariant.That moment invariants to establish 3D facial model with translation scale and rotation invariance based on preprocessing and normalization of a 3D facial model on.The facial expression classification application background by facial modeling realization of the use of multi dimension 3D-Zernike descriptors for intelligent support vector machine learning techniques to study the analysis and recognition system based on the analysis of recognition,optimization model based on D-ABC algorithm,improve the robustness of the model and the system reliability,and the use of the system for the classification and recognition of 3D facial data.
出处
《自动化技术与应用》
2015年第3期72-75,共4页
Techniques of Automation and Applications
关键词
三维不变矩
人脸识别技术
优化算法
three dimensional invariant moment
face recognition technology
optimization algorithm
