One shot ancient character recognition with siamese similarity network
Abstract Ancient character recognition is not only important for the study and understanding of ancient history but also has a profound impact on the inheritance and development of national culture. In order to reduce the study of difficult professional knowledge of ancient characters, and meanwhile...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2022-09-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-022-18986-z |
| _version_ | 1811274908419751936 |
|---|---|
| author | Xuxing Liu Weize Gao Rankang Li Yu Xiong Xiaoqin Tang Shanxiong Chen |
| author_facet | Xuxing Liu Weize Gao Rankang Li Yu Xiong Xiaoqin Tang Shanxiong Chen |
| author_sort | Xuxing Liu |
| collection | DOAJ |
| description | Abstract Ancient character recognition is not only important for the study and understanding of ancient history but also has a profound impact on the inheritance and development of national culture. In order to reduce the study of difficult professional knowledge of ancient characters, and meanwhile overcome the lack of data, class imbalance, diversification of glyphs, and open set recognition problems in ancient characters, we propose a Siamese similarity network based on a similarity learning method to directly learn input similarity and then apply the trained model to establish one shot classification task for recognition. Multi-scale fusion backbone structure and embedded structure are proposed in the network to improve the model's ability to extract features. We also propose the soft similarity contrast loss function for the first time, which ensures the optimization of similar images with higher similarity and different classes of images with greater differences while reducing the over-optimization of back-propagation leading to model overfitting. Specially, we propose a cumulative class prototype based on our network to solve the deviation problem of the mean class prototype and obtain a good class representation. Since new ancient characters can still be found in reality, our model has the ability to reject unknown categories while identifying new ones. A large number of experiments show that our proposed method has achieved high-efficiency discriminative performance and obtained the best performance over the methods of traditional deep learning and other classic one-shot learning. |
| first_indexed | 2024-04-12T23:28:40Z |
| format | Article |
| id | doaj.art-c87f11e35ce740839ded7369c707e4d5 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-04-12T23:28:40Z |
| publishDate | 2022-09-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-c87f11e35ce740839ded7369c707e4d52022-12-22T03:12:21ZengNature PortfolioScientific Reports2045-23222022-09-0112111510.1038/s41598-022-18986-zOne shot ancient character recognition with siamese similarity networkXuxing Liu0Weize Gao1Rankang Li2Yu Xiong3Xiaoqin Tang4Shanxiong Chen5College of Computer and Information Science, Southwest UniversityCollege of Computer and Information Science, Southwest UniversityCollege of Computer and Information Science, Southwest UniversityCollege of Computer and Information Science, Southwest UniversityCollege of Computer and Information Science, Southwest UniversityCollege of Computer and Information Science, Southwest UniversityAbstract Ancient character recognition is not only important for the study and understanding of ancient history but also has a profound impact on the inheritance and development of national culture. In order to reduce the study of difficult professional knowledge of ancient characters, and meanwhile overcome the lack of data, class imbalance, diversification of glyphs, and open set recognition problems in ancient characters, we propose a Siamese similarity network based on a similarity learning method to directly learn input similarity and then apply the trained model to establish one shot classification task for recognition. Multi-scale fusion backbone structure and embedded structure are proposed in the network to improve the model's ability to extract features. We also propose the soft similarity contrast loss function for the first time, which ensures the optimization of similar images with higher similarity and different classes of images with greater differences while reducing the over-optimization of back-propagation leading to model overfitting. Specially, we propose a cumulative class prototype based on our network to solve the deviation problem of the mean class prototype and obtain a good class representation. Since new ancient characters can still be found in reality, our model has the ability to reject unknown categories while identifying new ones. A large number of experiments show that our proposed method has achieved high-efficiency discriminative performance and obtained the best performance over the methods of traditional deep learning and other classic one-shot learning.https://doi.org/10.1038/s41598-022-18986-z |
| spellingShingle | Xuxing Liu Weize Gao Rankang Li Yu Xiong Xiaoqin Tang Shanxiong Chen One shot ancient character recognition with siamese similarity network Scientific Reports |
| title | One shot ancient character recognition with siamese similarity network |
| title_full | One shot ancient character recognition with siamese similarity network |
| title_fullStr | One shot ancient character recognition with siamese similarity network |
| title_full_unstemmed | One shot ancient character recognition with siamese similarity network |
| title_short | One shot ancient character recognition with siamese similarity network |
| title_sort | one shot ancient character recognition with siamese similarity network |
| url | https://doi.org/10.1038/s41598-022-18986-z |
| work_keys_str_mv | AT xuxingliu oneshotancientcharacterrecognitionwithsiamesesimilaritynetwork AT weizegao oneshotancientcharacterrecognitionwithsiamesesimilaritynetwork AT rankangli oneshotancientcharacterrecognitionwithsiamesesimilaritynetwork AT yuxiong oneshotancientcharacterrecognitionwithsiamesesimilaritynetwork AT xiaoqintang oneshotancientcharacterrecognitionwithsiamesesimilaritynetwork AT shanxiongchen oneshotancientcharacterrecognitionwithsiamesesimilaritynetwork |