Robust sclera recognition based on a local spherical structure

作者：

Highlights：

• A texture-based feature, called a vessel shape feature (VSF), is proposed to represent the shape of vessels in the sclera image.

• Inspired by the anatomical structure of the human eye, a local structure representation is devised to analyze the topological relationships of interest points in a 3D spherical space.

• By combining the VSF with the LSS to create a scheme called VSF-LSS, nonlinear deformation of the 2D sclera pattern can be addressed.

• An efficient coarse-to-fine algorithm is formulated for one-to-one matching, even if the numbers of interest points detected from any two sclera images are unequal.

摘要

•A texture-based feature, called a vessel shape feature (VSF), is proposed to represent the shape of vessels in the sclera image.•Inspired by the anatomical structure of the human eye, a local structure representation is devised to analyze the topological relationships of interest points in a 3D spherical space.•By combining the VSF with the LSS to create a scheme called VSF-LSS, nonlinear deformation of the 2D sclera pattern can be addressed.•An efficient coarse-to-fine algorithm is formulated for one-to-one matching, even if the numbers of interest points detected from any two sclera images are unequal.

论文关键词：Biometrics,Sclera recognition,Vessel shape feature (VSF),Local spherical structure (LSS),Local descriptor

论文评审过程：Received 19 August 2020, Revised 5 April 2021, Accepted 11 October 2021, Available online 15 October 2021, Version of Record 23 October 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.116081