Representation Learning Method for Circular Seal Based on Modified MLP-Mixer
This study proposes Stamp-MLP, an enhanced seal impression representation learning technique based on MLP-Mixer. Instead of using the patch linear mapping preprocessing method, this technique uses circular seal remapping, which reserves the seals’ underlying pixel-level information. In the proposed...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-11-01
|
| Series: | Entropy |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1099-4300/25/11/1521 |
| _version_ | 1827639894465314816 |
|---|---|
| author | Yuan Cao You Zhou Zhiwen Zhang Enyi Yao |
| author_facet | Yuan Cao You Zhou Zhiwen Zhang Enyi Yao |
| author_sort | Yuan Cao |
| collection | DOAJ |
| description | This study proposes Stamp-MLP, an enhanced seal impression representation learning technique based on MLP-Mixer. Instead of using the patch linear mapping preprocessing method, this technique uses circular seal remapping, which reserves the seals’ underlying pixel-level information. In the proposed Stamp-MLP, the average pooling is replaced by a global pooling of attention to extract the information more comprehensively. There were three classification tasks in our proposed method: categorizing the seal surface, identifying the product type, and distinguishing individual seals. The three tasks shared an identical dataset comprising 81 seals, encompassing 16 distinct seal surfaces, with each surface featuring six diverse product types. The experiment results showed that, in comparison to MLP-Mixer, VGG16, and ResNet50, the proposed Stamp-MLP achieved the highest classification accuracy (89.61%) in seal surface classification tasks with fewer training samples. Meanwhile, Stamp-MLP outperformed the others with accuracy rates of 90.68% and 91.96% in the product type and seal impression classification tasks, respectively. Moreover, Stamp-MLP had the fewest model parameters (2.67 M). |
| first_indexed | 2024-03-09T16:50:27Z |
| format | Article |
| id | doaj.art-056fda4effc543a8b1ec25d5910f33b3 |
| institution | Directory Open Access Journal |
| issn | 1099-4300 |
| language | English |
| last_indexed | 2024-03-09T16:50:27Z |
| publishDate | 2023-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Entropy |
| spelling | doaj.art-056fda4effc543a8b1ec25d5910f33b32023-11-24T14:41:00ZengMDPI AGEntropy1099-43002023-11-012511152110.3390/e25111521Representation Learning Method for Circular Seal Based on Modified MLP-MixerYuan Cao0You Zhou1Zhiwen Zhang2Enyi Yao3College of Information Science and Engineering, Hohai University, Changzhou 213022, ChinaCollege of Information Science and Engineering, Hohai University, Changzhou 213022, ChinaCollege of Information Science and Engineering, Hohai University, Changzhou 213022, ChinaSchool of Microelectronics, South China University of Technology, Guangzhou 511442, ChinaThis study proposes Stamp-MLP, an enhanced seal impression representation learning technique based on MLP-Mixer. Instead of using the patch linear mapping preprocessing method, this technique uses circular seal remapping, which reserves the seals’ underlying pixel-level information. In the proposed Stamp-MLP, the average pooling is replaced by a global pooling of attention to extract the information more comprehensively. There were three classification tasks in our proposed method: categorizing the seal surface, identifying the product type, and distinguishing individual seals. The three tasks shared an identical dataset comprising 81 seals, encompassing 16 distinct seal surfaces, with each surface featuring six diverse product types. The experiment results showed that, in comparison to MLP-Mixer, VGG16, and ResNet50, the proposed Stamp-MLP achieved the highest classification accuracy (89.61%) in seal surface classification tasks with fewer training samples. Meanwhile, Stamp-MLP outperformed the others with accuracy rates of 90.68% and 91.96% in the product type and seal impression classification tasks, respectively. Moreover, Stamp-MLP had the fewest model parameters (2.67 M).https://www.mdpi.com/1099-4300/25/11/1521seal recognitionMLP-Mixerrepresentation learning |
| spellingShingle | Yuan Cao You Zhou Zhiwen Zhang Enyi Yao Representation Learning Method for Circular Seal Based on Modified MLP-Mixer Entropy seal recognition MLP-Mixer representation learning |
| title | Representation Learning Method for Circular Seal Based on Modified MLP-Mixer |
| title_full | Representation Learning Method for Circular Seal Based on Modified MLP-Mixer |
| title_fullStr | Representation Learning Method for Circular Seal Based on Modified MLP-Mixer |
| title_full_unstemmed | Representation Learning Method for Circular Seal Based on Modified MLP-Mixer |
| title_short | Representation Learning Method for Circular Seal Based on Modified MLP-Mixer |
| title_sort | representation learning method for circular seal based on modified mlp mixer |
| topic | seal recognition MLP-Mixer representation learning |
| url | https://www.mdpi.com/1099-4300/25/11/1521 |
| work_keys_str_mv | AT yuancao representationlearningmethodforcircularsealbasedonmodifiedmlpmixer AT youzhou representationlearningmethodforcircularsealbasedonmodifiedmlpmixer AT zhiwenzhang representationlearningmethodforcircularsealbasedonmodifiedmlpmixer AT enyiyao representationlearningmethodforcircularsealbasedonmodifiedmlpmixer |