Hierarchical kernel-based rotation and scale invariant similarity

作者：

Highlights：

• An effective rotation and scale invariant similarity measure is proposed.

• The supervised template sets construction method is based on the sparse principle.

• The proposed method can lead to high classification accuracy and is fast.

摘要

Highlights•An effective rotation and scale invariant similarity measure is proposed.•The supervised template sets construction method is based on the sparse principle.•The proposed method can lead to high classification accuracy and is fast.

论文关键词：Image similarity measure,Log-polar transform,Rotation and scale invariance,Hierarchical kernels

论文评审过程：Received 6 January 2013, Revised 12 August 2013, Accepted 7 October 2013, Available online 26 October 2013.

论文官网地址：https://doi.org/10.1016/j.patcog.2013.10.008