节点文献

基于差图像的人脸动态表情特征提取

Facial Dynamic Feature Extraction of Human Being Based on Different Images

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 何显文左坤隆

【Author】 HE Xian-wen1,ZUO Kun-long2(1.School of Maths & Computer,Gannan Normal University,Ganzhuo 341000,China;2.Tianjin University,Tianjin 300072,China)

【机构】 赣南师范学院数学与计算机科学学院天津大学

【摘要】 在人脸表情的识别中,动态表情图像包含丰富的信息,应用非常广泛.在动态表情特征的提取中,差图像特征提取是通过基于逆向组合的图像对齐方法实现的,图像对齐方法可以在一定程度上避免差图像特征与几何特征的误差相关性,同时逆向组合又可以保证在不损失图像对齐精度的前提下大大提高运算速度.

【Abstract】 During the course of facial expression recognition,dynamic expression image includes fruitful information and is used widely.During the extraction of dynamic features of expression,the extraction of different images can be realized by reversing compositional method of image alignment,the method of image alignment can avoid the conflict between the features in different image and geometric features.Moreover,reverse composition will double the calculating speed without doing harm to the images.

  • 【文献出处】 赣南师范学院学报 ,Journal of Gannan Normal University , 编辑部邮箱 ,2008年06期
  • 【分类号】TP391.41
  • 【被引频次】3
  • 【下载频次】106
节点文献中: 

本文链接的文献网络图示:

本文的引文网络