计算机与现代化

• 图像处理 • 上一篇    下一篇

基于双线性模型的人脸表情识别技术

  

  1. (北京工业大学,北京 100124)
  • 收稿日期:2014-03-25 出版日期:2014-07-16 发布日期:2014-07-17
  • 作者简介:徐欢(1987-),男,吉林长春人,北京工业大学硕士研究生,研究方向:表情识别。

Facial Expression Recognition Techniques Based on Bilinear Model

  1. (Beijing University of Technology, Beijing 100124, China)
  • Received:2014-03-25 Online:2014-07-16 Published:2014-07-17

摘要: 针对现阶段人脸表情识别过程中所遇到的问题,基于三维数据库BU-3DFE中的三维表情数据,研究三维人脸表情数据的点云对齐及基于对齐数据的双线性模型建立,对基于双线性模型的识别算法加以改进,形成新的识别分类算法,降低原有算法中身份特征参与计算的比重,最大可能地降低身份特征对于整个表情识别过程的影响。旨在提高表情识别的结果,最终实现高鲁棒性的三维表情识别。

关键词: 面部表情识别, 特征提取, 双线性模型, TPS对齐

Abstract: Aiming at the problems existing in facial expression recognition currently, based on the data in the 3D expression database BU-3DFE, we study the point cloud alignment of 3D facial expression data, establish the bilinear models based on the alignment data, and improve the recognition algorithms based on bilinear model in order to form the new recognition and classification algorithms, to reduce the quantity of identity feature calculation in original algorithm, to minimize the influence of identity feature on the total expression recognition process, to improve the results of facial expression recognition, and to ultimately achieve the high robustness of 3D facial expression recognition.

Key words: facial expression recognition, feature extraction, bilinear model, TPS alignment

中图分类号: