基于ResNet和迁移学习的古印章文本识别

doi:10.3778/j.issn.1002-8331.2101-0247

摘要/Abstract

摘要： 古印章文本因图像退化与超多分类等特点导致识别难度大，部分字符的标注数据不足造成基于深度学习的模型识别准确率不高，泛化能力差。针对上述问题，提出基于深度残差网络（ResNet）和迁移学习的古印章文本识别方法。使用深度残差网络作为特征提取网络，利用人工合成字符样本作为源域进行预训练。将自建古印章文本识别数据集作为目标域，引入迁移学习并结合数据增强和标签平滑策略建立分类模型。最后，对比多种网络下的识别结果并验证迁移学习有效性。结果表明，该方法可以有效提升识别准确率。

关键词: 古印章文本, 文本识别, 深度残差网络, 迁移学习

Abstract: It is difficult to recognize historical Chinese seal text due to image degradation and super classification. In addition, the insufficient annotation data lead to poor generalization ability and classification accuracy. According to the above problems, a historical Chinese seal text recognition method based on ResNet and transfer learning is proposed. Firstly, using synthetic dataset as source domain, a pre-trained model is trained on deep residual network. Secondly, transfer learning is introduced to model combining data enhancement and label smoothing. In this process, the historical Chinese seal text dataset is taken as target domain. Finally, the recognition results in different networks are compared and the transfer learning effectiveness is analyzed. The experimental results show that this method can improve recognition accuracy effectively.

Key words: historical Chinese seal text, text recognition, residual network, transfer learning

陈娅娅, 刘全香, 王凯丽, 易尧华. 基于ResNet和迁移学习的古印章文本识别[J]. 计算机工程与应用, 2022, 58(10): 125-131.

CHEN Yaya, LIU Quanxiang, WANG Kaili, YI Yaohua. Historical Chinese Seal Text Recognition Based on ResNet and Transfer Learning[J]. Computer Engineering and Applications, 2022, 58(10): 125-131.

参考文献

[1] WANG B，WEI W Y，HU H T，et al.Hybrid model for Chinese character recognition based on Tesseract-OCR[J].International Journal of Internet Protocol Technology，2020，13（2）：102.
[2] HUANG T C，ZHANG S Y，DUAN X，et al.On-line handy handwriting Chinese characters input for non-Chinese speakers based on wavelet neural network[J].International Journal of Pattern Recognition and Artificial Intelligence，2016，30（6）：1-16.
[3] 杨恒杰.自然场景下文字识别与信息抽取的研究与应用[D].厦门：华侨大学，2020.
YANG H J.Research and application of character recognition and information extraction in natural scenes[D].Xiamen：Huaqiao University，2020.
[4] GUO P C，WANG J Q，LI Y，et al.Fast Chinese calligraphic character recognition with large-scale data[J].Multimedia Tools & Applications，2015，74（17）：7221-7238.
[5] CLANUWAT T，LAMB A，KITAMOTO A.KuroNet：pre-modern Japanese Kuzushiji character recognition with deep learning[C]//2019 International Conference on Document Analysis and Recognition，2019.
[6] SUN B，HUA S，LI S T，et al.Graph-matching-based character recognition for Chinese seal images[J].Science China（Information Sciences），2019，62（9）：33-46.
[7] DAI T，SUN B.Novel features for character extraction of historical Chinese seal images[J].Sensing and Imaging，2019，20：32.
[8] WANG Z，LIAN J，SONG C，et al.CSRS：a Chinese seal recognition system with multi-task learning and automatic background generation[J].IEEE Access，2019，7：96628-96638.
[9] 张顺，龚怡宏，王进军.深度卷积神经网络的发展及其在计算机视觉领域的应用[J].计算机学报，2019，42（3）：453-482.
ZHANG S，GONG Y H，WANG J J.The development of deep convolution neural network and its application on computer vision[J].Chinese Journal of Computer，2019，42（3）：453-482.
[10] KRIZHEVSKY A，SUTSKEVER I，HINTON G E.ImageNet classification with deep convolutional neural networks[C]//Advances in Neural Information Processing Systems，2012：1106-1114.
[11] SIMONYAN K，ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[J].arXiv：1409.
1556，2014.
[12] SZEGEDY C，LIU W，JIA Y，et al.Going deeper with convolutions[C]//2015 IEEE Conference on Computer Vision and Pattern Recognition，2015：1-9.
[13] HE K，ZHANG X，REN S，et al.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition，2016：770-778.
[14] ZHONG Z，JIN L，FENG Z.Multi-font printed Chinese character recognition using multi-pooling convolutional neural network[C]//13th International Conference on Document Analysis and Recognition，2015：96-100.
[15] WANG K L，YI Y H，LIU J，et al.Multi-scene ancient Chinese text recognition[J].Neurocomputing，2020，377：64-72.
[16] PAN J S，YANG Q.A survey on transfer learning[J].IEEE Transactions on Knowledge and Data Engineering，2010，22（10）：1345-1359.
[17] 姚可欣，曹卫群.Trans-Net：基于迁移学习的手写简笔画识别[J].计算机工程与应用，2021，57（3）：182-188.
YAO K X，CAO W Q.Trans-Net：stick figure recognition based on transfer learning[J].Computer Engineering and Applications，2021，57（3）：182-188.
[18] OTSU N.A threshold selection method from gray level histograms[J].IEEE Transactions on System，Man and Cybernetic，1979，9（1）：62-66.
[19] HE K，ZHANG X，REN S，et al.Identity mappings in deep residual networks[C]//14th European Conference on Computer Vision.Cham：Springer，2016.
[20] MOLCHANOV P，TYREE S，KARRAS T，et al.Pruning convolutional neural networks for resource efficient inference[C]//5th International Conference on Learning Representations，2017.
[21] ABADI M，AGARWAL A，BARHAM P，et al.TensorFlow：large-scale machine learning on heterogeneous distributed systems[J].arXiv：1603.04467，2016.