MBRNet：融合残差连接的多分支手写字符识别网络

doi:10.3778/j.issn.1002-8331.2308-0161

摘要/Abstract

摘要： 脱机手写中文字符识别（handwritten Chinese character recognition，HCCR）在计算机视觉领域一直是一个巨大的挑战。相比传统方法，基于深度学习的网络通过训练大量数据在识别任务中取得了差异化的效果，但识别效果依旧处于发展过程中。基于此，结合DW卷积和残差连接设计了一种多分支残差模块，该模块通过DW卷积以较小的内存和参数量为代价来加深网络深度，增强网络的特征提取能力；再通过残差连接抑制网络梯度问题和退化问题；另外，提出了一种多分支权重算法，来改善多分支残差模块中各分支的权重分配问题；并将六个以多分支残差模块为主的结构线性连接，组成HCCR识别网络。该模型在CASIA-HWDB1.0、CASIA-HWDB1.1、ICDAR2013数据集上的识别准确率分别达到了97.77%、97.30%、97.64%，表现出高精度的识别效果。

关键词: 手写中文字符识别（HCCR）, 多分支残差模块, DW卷积, 残差连接, 多分支权重

Abstract: Offline handwritten Chinese character recognition (HCCR) has been a great challenge in the field of computer vision. Compared with traditional methods, deep learning-based networks have achieved differentiated results in the recognition task by training a large amount of data, but the recognition effect is still in the process of development. Based on this, a multi-branch residual block is designed by combining DW convolution operations and residual connections. In this block, DW convolution operations increase the depth of the network and enhance feature extraction capabilities at the cost of smaller memory usage and parameter count. And the residual connections facilitate data spiraling flow, effectively mitigating gradient and degradation issues in the network. Furthermore, a multi-branch weight algorithm is proposed to address the weight allocation issue for the branches within the multi-branch residual block. Six multi-branch residual blocks are linearly connected to form the HCCR recognition network. The model achieves recognition accuracies of 97.77%, 97.30%, and 97.64% on the CASIA-HWDB1.0, CASIA-HWDB1.1, and ICDAR2013 datasets, respectively, showing high recognition accuracy.

Key words: handwritten Chinese character recognition (HCCR), multi-branch residual block, DW convolution, residual connection, multi-branch weighting

李钢, 陈太兵, 杨之博, 范屹, 张玲. MBRNet：融合残差连接的多分支手写字符识别网络[J]. 计算机工程与应用, 2024, 60(24): 149-157.

LI Gang, CHEN Taibing, YANG Zhibo, FAN Yi, ZHANG Ling. MBRNet：Multi-Branch Handwritten Character Recognition Network with Integrated Residual Connection[J]. Computer Engineering and Applications, 2024, 60(24): 149-157.

参考文献

[1] LIU C L, JAEGER S, NAKAGAWA M. Online recognition of Chinese characters: the state-of-the-art[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2004, 26(2): 198-213.
[2] LIU B, XU X, ZHANG Y. Offline handwritten Chinese text recognition with convolutional neural networks[J]. arXiv:2006.
15619, 2020.
[3] DAI R, LIU C, XIAO B. Chinese character recognition: history, status and prospects[J]. Frontiers of Computer Science in China, 2007, 1: 126-136.
[4] LIU C L, KOGA M, FUJISAWA H. Lexicon-driven segmentation and recognition of handwritten character strings for Japanese address reading[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(11): 1425-1437.
[5] KIMURA F, TAKASHINA K, TSURUOKA S, et al. Modified quadratic discriminant functions and the application to Chinese character recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1987(1): 149-153.
[6] JIN L, HUANG J, YIN J, et al. Deformation transformation for handwritten Chinese character shape correction[C]//Proceedings of the Third International Conference on Advances in Multimodal Interfaces, Beijing, China, October 14-16, 2000. Berlin, Heidelberg: Springer, 2000: 450-457.
[7] LONG T, JIN L. Building compact MQDF classifier for large character set recognition by subspace distribution sharing[J]. Pattern Recognition, 2008, 41(9): 2916-2925.
[8] LIU C, YIN F, WANG D, et al. Online and offline handwritten Chinese character recognition: benchmarking on new databases[J]. Pattern Recognition, 2013, 46(1): 155-162.
[9] ZHANG X Y, BENGIO Y, LIU C. Online and offline handwritten Chinese character recognition: a comprehensive study and new benchmark[J]. Pattern Recognition, 2017, 61: 348-360.
[10] WANG Z, YU Y, WANG Y, et al. Robust end-to-end offline Chinese handwriting text page spotter with text kernel[C]// Proceedings of the International Conference on Document Analysis and Recognition (ICDAR 2021) , Lausanne, Switzerland, September 5-10, 2021: 21-35.
[11] LIU C L, SAKO H, FUJISAWA H. Discriminative learning quadratic discriminant function for handwriting recognition[J]. IEEE Transactions on Neural Networks, 2004, 15(2): 430-444.
[12] LIU C L, YIN F, WANG D H, et al. Chinese handwriting recognition contest 2010[C]//Proceedings of the 2010 Chinese Conference on Pattern Recognition (CCPR), 2010: 1-5.
[13] LIU C L, YIN F, WANG Q F, et al. ICDAR 2011 Chinese handwriting recognition competition[C]//Proceedings of the International Conference on Document Analysis and Recognition, 2011: 1464-1469.
[14] YIN F, WANG Q F, ZHANG X Y, et al. ICDAR 2013 Chinese handwriting recognition competition[C]//Proceedings of the 2013 12th International Conference on Document Analysis and Recognition, 2013: 1464-1470.
[15] CIRESAN D C, MEIER U, MASCI J, et al. Flexible, high performance convolutional neural networks for image classification[C]//Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence, 2011.
[16] BENGIO Y, COURVILLE A, VINCENT P. Representation learning: a review and new perspectives[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(8): 1798-1828.
[17] LECUN Y, BOTTOU L, BENGIO Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324.
[18] LECUN Y, BOSER B, DENKER J, et al. Backpropagation applied to handwritten zip code recognition[J]. Neural Computation, 1989, 1(4): 541-551.
[19] STEINKRAUS D, BUCK I, SIMARD P Y. Using GPUs for machine learning algorithms[C]//Proceedings of the Eighth International Conference on Document Analysis and Recognition (ICDAR’05), 2005.
[20] ZHONG Z Y, JIN L W, XIE Z C. High performance offline handwritten Chinese character recognition using GoogLeNet and directional feature maps[C]//Proceedings of the 2015 13th International Conference on Document Analysis and Recognition(ICDAR), Nancy, France, 2015: 846-850.
[21] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Advances in Neural Information Processing Systems, 2017.
[22] DAN Y, ZHU Z, JIN W, et al. S-Swin Transformer: simplified Swin Transformer model for offline handwritten Chinese character recognition[J]. PeerJ Computer Science, 2022, 8: e1093.
[23] GENG S, ZHU Z, WANG Z, et al. LW-ViT: the lightweight vision transformer model applied in offline handwritten Chinese character recognition[J]. Electronics, 2023, 12(7): 1693.
[24] WU C, FAN W, HE Y, et al. Handwritten character recognition by alternately trained relaxation convolutional neural network[C]//Proceedings of the 2014 14th International Conference on Frontiers in Handwriting Recognition, 2014: 291-296.
[25] ZHONG Z, JIN L, XIE Z. High performance offline handwritten Chinese character recognition using GoogLeNet and directional feature maps[C]//Proceedings of the 2015 13th International Conference on Document Analysis and Recognition (ICDAR), 2015.
[26] XIAO X, JIN L, YANG Y, et al. Building fast and compact convolutional neural networks for offline handwritten Chinese character recognition[J]. arXiv:1702.07975, 2017.
[27] 黄婉蓉, 何凯, 刘坤, 等. 基于注意力机制的手写体中文字符识别[J]. 激光与光电子学进展, 2020, 57(8): 37-42.
HUANG W R, HE K, LIU K, et al. Handwritten Chinese character recognition based on attention mechanism[J]. Laser & Optoelectronics Progress, 2020, 57(8): 37-42.
[28] MELNYK P, YOY Z, LI K. A high-performance CNN method for offline handwritten Chinese character recognition and visualization[J]. Soft Computing, 2019, 24: 7977-7987.
[29] LI Z, TENG N, JIN M, et al. Building efficient CNN architecture for offline handwritten Chinese character recognition[J]. International Journal on Document Analysis and Recognition (IJDAR), 2018, 21: 233-240.