Covariance descriptors on a Gaussian manifold and their application to image set classification

作者：

Highlights：

• Characterizing the similarities between the regions that contain more comprehensive information than pixels.

• Presenting two methods for computing Riemannian local difference vector on Gaussian manifold (RieLDV-G) and using RieLDV-G to define deviation.

• Providing a novel framework for computing covariance on Gaussian manifold and generating the proposed Riemannian covariance descriptors (RieCovDs).

摘要

•Characterizing the similarities between the regions that contain more comprehensive information than pixels.•Presenting two methods for computing Riemannian local difference vector on Gaussian manifold (RieLDV-G) and using RieLDV-G to define deviation.•Providing a novel framework for computing covariance on Gaussian manifold and generating the proposed Riemannian covariance descriptors (RieCovDs).

论文关键词：Covariance descriptors,Riemannian local difference vector,Riemannian covariance descriptors,Image set classification

论文评审过程：Received 14 May 2019, Revised 15 May 2020, Accepted 16 May 2020, Available online 19 May 2020, Version of Record 17 June 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107463