Discriminative quadratic feature learning for handwritten Chinese character recognition

作者：

Highlights：

• We propose a feature learning method for handwritten Chinese character recognition.

• Quadratic correlation between original gradient features is utilized.

• Discriminative learning guarantees the discriminability of quadratic features.

• The proposed method outperforms deep convolutional neural networks with much faster test speed.

摘要

Highlights•We propose a feature learning method for handwritten Chinese character recognition.•Quadratic correlation between original gradient features is utilized.•Discriminative learning guarantees the discriminability of quadratic features.•The proposed method outperforms deep convolutional neural networks with much faster test speed.

论文关键词：Handwritten Chinese character recognition,Discriminative feature learning,Quadratic correlation,Dimensionality promotion,Training set expansion

论文评审过程：Received 7 February 2015, Revised 26 June 2015, Accepted 22 July 2015, Available online 1 August 2015, Version of Record 28 September 2015.

论文官网地址：https://doi.org/10.1016/j.patcog.2015.07.007