首页> 外文期刊>Image and Vision Computing >An SE(3) invariant description for 3D face recognition
【24h】

An SE(3) invariant description for 3D face recognition

机译:3D人脸识别的SE(3)不变描述

获取原文
获取原文并翻译 | 示例
           

摘要

Here, we intend to introduce a novel 3D face description which is invariant under the Special Euclidean group SE(3) and independent to the original surface parameterization. It is well known that it is too difficult to define a relative invariant parameterization of a general curved surface. In the present work, we introduce the multi-polar geodesic representation of R-3 surfaces. It allows to reach an isotropic canonical parameterization relatively to SE(3) due to the fact that the face surfaces can be easily assumed to be a graph of a function from R-2 to R. The principal curvature fields according to a three-polar parameterization are considered. A statistical study is made to choose a good configuration of reference points. The performances of the novel description are tested on the standard database FRGC v2.0. Many recognition scenarios are established. The obtained results are very competitive with the state of the art. (C) 2019 Elsevier B.V. All rights reserved.
机译:在这里,我们打算介绍一种新颖的3D人脸描述,它在特殊欧几里得组SE(3)下是不变的,并且与原始表面参数化无关。众所周知,定义一般曲面的相对不变的参数化太困难了。在当前的工作中,我们介绍了R-3表面的多极测地线表示。由于可以很容易地假设端面是从R-2到R的函数的图,因此它可以实现相对于SE(3)的各向同性规范参数化。根据三极点的主曲率场考虑参数化。进行统计研究以选择参考点的良好配置。新颖描述的性能在标准数据库FRGC v2.0上进行了测试。建立了许多识别方案。所获得的结果与现有技术非常有竞争力。 (C)2019 Elsevier B.V.保留所有权利。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号