Self-supervised few-shot learning for real-time traffic sign classification
Although supervised approaches for traffic sign classification have demonstrated excellent performance, they are limited to classifying several traffic signs defined in the training dataset. This prevents them from being applied to different domains, i.e., different countries. Herein, we propose a s...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Universitas Ahmad Dahlan
2024-02-01
|
| Series: | IJAIN (International Journal of Advances in Intelligent Informatics) |
| Subjects: | |
| Online Access: | http://ijain.org/index.php/IJAIN/article/view/1522 |
| _version_ | 1827323343442804736 |
|---|---|
| author | Anh-Khoa Tho Nguyen Tin Tran Phuc Hong Nguyen Vinh Quang Dinh |
| author_facet | Anh-Khoa Tho Nguyen Tin Tran Phuc Hong Nguyen Vinh Quang Dinh |
| author_sort | Anh-Khoa Tho Nguyen |
| collection | DOAJ |
| description | Although supervised approaches for traffic sign classification have demonstrated excellent performance, they are limited to classifying several traffic signs defined in the training dataset. This prevents them from being applied to different domains, i.e., different countries. Herein, we propose a self-supervised approach for few-shot learning-based traffic sign classification. A center-awareness similarity network is designed for the traffic sign problem and trained using an optical flow dataset. Unlike existing supervised traffic sign classification methods, the proposed method does not depend on traffic sign categories defined by the training dataset. It applies to any traffic signs from different countries. We construct a Korean traffic sign classification (KTSC) dataset, including 6000 traffic sign samples and 59 categories. We evaluate the proposed method with baseline methods using the KTSC, German traffic sign, and Belgian traffic sign classification datasets. Experimental results show that the proposed method extends the ability of existing supervised methods and can classify any traffic sign, regardless of region/country dependence. Furthermore, the proposed approach significantly outperforms baseline methods for patch similarity. This approach provides a flexible and robust solution for classifying traffic signs, allowing for accurate categorization of every traffic sign, regardless of regional or national differences. |
| first_indexed | 2024-04-25T01:43:26Z |
| format | Article |
| id | doaj.art-a1774560c82646ecbacc5dc0cc726883 |
| institution | Directory Open Access Journal |
| issn | 2442-6571 2548-3161 |
| language | English |
| last_indexed | 2024-04-25T01:43:26Z |
| publishDate | 2024-02-01 |
| publisher | Universitas Ahmad Dahlan |
| record_format | Article |
| series | IJAIN (International Journal of Advances in Intelligent Informatics) |
| spelling | doaj.art-a1774560c82646ecbacc5dc0cc7268832024-03-08T03:14:06ZengUniversitas Ahmad DahlanIJAIN (International Journal of Advances in Intelligent Informatics)2442-65712548-31612024-02-0110117118510.26555/ijain.v10i1.1522288Self-supervised few-shot learning for real-time traffic sign classificationAnh-Khoa Tho Nguyen0Tin Tran1Phuc Hong Nguyen2Vinh Quang Dinh3Department of Computer Science, Vietnamese German UniversityAI Graduate School, Gwangju Institute of Science and TechnologyDepartment of Software Engineering, Eastern International UniversityDepartment of Computer Science, Vietnamese German UniversityAlthough supervised approaches for traffic sign classification have demonstrated excellent performance, they are limited to classifying several traffic signs defined in the training dataset. This prevents them from being applied to different domains, i.e., different countries. Herein, we propose a self-supervised approach for few-shot learning-based traffic sign classification. A center-awareness similarity network is designed for the traffic sign problem and trained using an optical flow dataset. Unlike existing supervised traffic sign classification methods, the proposed method does not depend on traffic sign categories defined by the training dataset. It applies to any traffic signs from different countries. We construct a Korean traffic sign classification (KTSC) dataset, including 6000 traffic sign samples and 59 categories. We evaluate the proposed method with baseline methods using the KTSC, German traffic sign, and Belgian traffic sign classification datasets. Experimental results show that the proposed method extends the ability of existing supervised methods and can classify any traffic sign, regardless of region/country dependence. Furthermore, the proposed approach significantly outperforms baseline methods for patch similarity. This approach provides a flexible and robust solution for classifying traffic signs, allowing for accurate categorization of every traffic sign, regardless of regional or national differences.http://ijain.org/index.php/IJAIN/article/view/1522traffic sign classificationone-shot learningfew-shot learningself-supervised learningclip-based approach |
| spellingShingle | Anh-Khoa Tho Nguyen Tin Tran Phuc Hong Nguyen Vinh Quang Dinh Self-supervised few-shot learning for real-time traffic sign classification IJAIN (International Journal of Advances in Intelligent Informatics) traffic sign classification one-shot learning few-shot learning self-supervised learning clip-based approach |
| title | Self-supervised few-shot learning for real-time traffic sign classification |
| title_full | Self-supervised few-shot learning for real-time traffic sign classification |
| title_fullStr | Self-supervised few-shot learning for real-time traffic sign classification |
| title_full_unstemmed | Self-supervised few-shot learning for real-time traffic sign classification |
| title_short | Self-supervised few-shot learning for real-time traffic sign classification |
| title_sort | self supervised few shot learning for real time traffic sign classification |
| topic | traffic sign classification one-shot learning few-shot learning self-supervised learning clip-based approach |
| url | http://ijain.org/index.php/IJAIN/article/view/1522 |
| work_keys_str_mv | AT anhkhoathonguyen selfsupervisedfewshotlearningforrealtimetrafficsignclassification AT tintran selfsupervisedfewshotlearningforrealtimetrafficsignclassification AT phuchongnguyen selfsupervisedfewshotlearningforrealtimetrafficsignclassification AT vinhquangdinh selfsupervisedfewshotlearningforrealtimetrafficsignclassification |