GR-RNN: Global-context residual recurrent neural networks for writer identification

作者：

Highlights：

• We present a global-context residual recurrent neural network for writer identification.

• The proposed method integrates the global context and the information of a sequence of local fragments by recurrent neural networks.

• We evaluate the proposed method on four public datasets, showing state-of-the-art performance.

摘要

•We present a global-context residual recurrent neural network for writer identification.•The proposed method integrates the global context and the information of a sequence of local fragments by recurrent neural networks.•We evaluate the proposed method on four public datasets, showing state-of-the-art performance.

论文关键词：Writer identification,Recurrent neural network,Residual network,Local and global features

论文评审过程：Received 20 July 2020, Revised 5 March 2021, Accepted 31 March 2021, Available online 20 April 2021, Version of Record 3 May 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.107975