Discriminative quadratic feature learning for handwritten Chinese character recognition
Discriminative quadratic feature learning for handwritten Chinese character recognition
复制标题
手写汉字识别的判别二次特征学习
DOI：
10.1016/j.patcog.2015.07.007
复制
发表时间：
2016
影响因子：
8
通讯作者：
Cheng-Lin Liu
中科院分区：
文献类型：
--
作者：
Ming-Ke Zhou;Xu-Yao Zhang;Fei Yin;Cheng-Lin Liu
提出了一种用于手写汉字识别的特征学习方法，称为判别二次特征学习(DQFL)。在原有的梯度方向特征表示的基础上，利用特征间的二次相关来提升特征的维度，然后利用判别特征提取(DFE)进行降维。通过维提升和降维相结合的方法，我们可以学习到更具区分性和非线性的特征表示，从而可以显著提高分类精度。对于维度提升，利用了两种类型的相关性，即统计相关性和空间相关性。对字符图像不同区域的多个局部特征向量进行统计相关性计算，而空间相关性对两个位置特征之间的相关性进行编码。特征相关性使维度增加了40,000多个。然后，DFE将维度降低到300以下，而不会失去可区分性。使用最近原型分类器(NPC)、改进的二次判别函数(MQDF)和判别学习二次判别函数(DLQDF)进行分类。在CASIA-HWDB1.1标准数据集上的实验中，DQFL方法将NPC、MQDF和DLQDF的测试精度分别提高了4.94%、1.83%和1.82%。通过对训练集进行扩展，进一步提高了测试精度。在ICDAR 2013中文手写识别竞赛数据集上，DQFL+DLQDF分类器的性能优于基于深度卷积神经网络(CNN)的最佳参赛系统，且测试速度快得多。
In this paper, we propose a feature learning method for handwritten Chinese character recognition (HCCR), called discriminative quadratic feature learning (DQFL). Based on original gradient direction feature representation, quadratic correlation between features is used to promote the feature dimensionality, then discriminative feature extraction (DFE) is used for dimensionality reduction. By combining dimensionality promotion and reduction, we can learn a much more discriminative and nonlinear feature representation, which can then boost the classification accuracy significantly. For dimensionality promotion, two types of correlation are exploited, namely, statistical correlation and spatial correlation. Statistical correlation is computed on multiple local feature vectors in different regions of the character image; while spatial correlation encodes the dependency between features of two positions. Feature correlation increases the dimensionality by over 40,000. DFE then reduces the dimensionality to less than 300 without losing discriminability. Classification is performed using nearest prototype classifier (NPC), modified quadratic discriminant function (MQDF) and discriminative learning quadratic discriminant function (DLQDF). In experiments on the CASIA-HWDB1.1 standard dataset, the proposed DQFL method improves the test accuracies of NPC, MQDF and DLQDF by 4.94%, 1.83%, and 1.82%, respectively. The test accuracy is further improved by training set expansion. On the ICDAR 2013 Chinese handwriting recognition competition dataset, the proposed DQFL+DLQDF classifier outperforms the best participating system based on deep convolutional neural network (CNN), while the test speed is much faster.
登录
查看更多内容
DOI：
10.1016/0031-3203(90)90110-7
发表时间：
1990-09
期刊：
Pattern Recognit.
影响因子：
--
作者：
Hiromitsu Yamada;Kazuhiko Yamamoto;Taiichi Saito
通讯作者：
Hiromitsu Yamada;Kazuhiko Yamamoto;Taiichi Saito
DOI：
10.1016/s0031-3203(00)00018-2
发表时间：
2001-03
期刊：
Pattern Recognit.
影响因子：
--
作者：
Cheng-Lin Liu;M. Nakagawa
通讯作者：
Cheng-Lin Liu;M. Nakagawa
DOI：
10.1016/c2009-0-27872-x
发表时间：
1972
期刊：
影响因子：
--
作者：
H. Shimodaira;Iain Murray
通讯作者：
Iain Murray
DOI：
10.1002/0470854774.ch1
发表时间：
2006
期刊：
--
影响因子：
--
作者：
P. Pudil;P. Somol;M. Haindl
通讯作者：
P. Pudil;P. Somol;M. Haindl
DOI：
10.1109/icdar.1997.620551
发表时间：
1997-06
期刊：
Proceedings of the Fourth International Conference on Document Analysis and Recognition
影响因子：
--
作者：
T. Horiuchi;R. Haruki;Hiromitsu Yamada;Kazuhiko Yamamoto
通讯作者：
T. Horiuchi;R. Haruki;Hiromitsu Yamada;Kazuhiko Yamamoto