GLMNet: Graph learning-matching convolutional networks for feature matching

作者：

Highlights：

• Our model integrates graph learning and graph matching together in a unified network.

• We employ a Laplacian sharpening convolution to generate more discriminative node embeddings for matching task.

• We design a new constraint regularized loss to encode the one-to-one matching constraints.

摘要

•Our model integrates graph learning and graph matching together in a unified network.•We employ a Laplacian sharpening convolution to generate more discriminative node embeddings for matching task.•We design a new constraint regularized loss to encode the one-to-one matching constraints.

论文关键词：Graph matching,Graph learning,Graph convolutional network,Laplacian sharpening

论文评审过程：Received 27 January 2021, Revised 6 May 2021, Accepted 4 July 2021, Available online 15 July 2021, Version of Record 31 July 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108167