SQ-ViT: A Multi-Scale Vision Transformer With Quaternion for Endoscopic Images Classification
IEEE Transactions on Consumer Electronics · 2025-02 · 71 (1) · 828-838
DOI: 10.1109/tce.2024.3518755 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Consumer Electronics, 2025-02. Verfasst von Zhanjun Jin und 7 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 50 Literaturangaben. Gefördert von: Science and Technology Development Fund of Macau SAR, Key Areas Research and Development Program of Guangzhou, Guangdong Provincial Key Laboratory of Cyber-Physical System.
Abstract (Originalsprache)
In the field of medical consumer electronics, endoscopic imaging technology especially electronic nasopharyngoscope imaging, often suffers from low resolution, which poses a difficulty for endoscopic images classification due to the loss of image details. Recent advancements in Vision Transformer (ViT) based methods have shown promise in addressing this problem. However, ViT relies heavily on global context information to maintain performance, and the limited pixel count in low-resolution images poses a challenge in capturing adequate global context information. To address these challenges, we propose the Sequential Quaternion Vision Transformer (SQ-ViT), which improves multi-scale feature utilization by feeding sampled features into the subsequent encoder layers. Specifically, we introduce the Multi-scale Visual Feature Fusion (MVFF) module, which segments the image into multiple superpixel blocks and refines the contour and color information of the processed image, which helps to enhance the representation of visual features. Additionally, visual information would be captured more effectively by our proposed Quaternion Interactive Encoder (QIE). Experiments demonstrate the effectiveness of SQ-ViT in improving multi-scale feature utilization and addressing challenges in low-resolution endoscopic imaging for endoscopic images classification. The source code will be released at https://github.com/jinzhanjun625/SQ-ViT.
Förderung
- Science and Technology Development Fund of Macau SAR
- Key Areas Research and Development Program of Guangzhou
- Guangdong Provincial Key Laboratory of Cyber-Physical System
- Basic and Applied Research Foundation of Guangdong Province
Zitieren
APA
Jin, Z., Huang, G., Zhang, F., Yuan, X., Zhu, D., Tan, Z., Pun, C., & Zhong, G. (2025). SQ-ViT: A Multi-Scale Vision Transformer With Quaternion for Endoscopic Images Classification. IEEE Transactions on Consumer Electronics, 71(1), 828-838. https://doi.org/10.1109/tce.2024.3518755
Vancouver
Jin Z, Huang G, Zhang F, Yuan X, Zhu D, Tan Z, et al. SQ-ViT: A Multi-Scale Vision Transformer With Quaternion for Endoscopic Images Classification. IEEE Transactions on Consumer Electronics. 2025;71(1):828-838. doi:10.1109/tce.2024.3518755
BibTeX
@article{Jin2025,
author = {Jin, Zhanjun and Huang, Guoheng and Zhang, Feng and Yuan, Xiaochen and Zhu, Dingzhou and Tan, Zhe and Pun, Chi-Man and Zhong, Guo},
title = {SQ-ViT: A Multi-Scale Vision Transformer With Quaternion for Endoscopic Images Classification},
journal = {IEEE Transactions on Consumer Electronics},
year = {2025},
volume = {71},
number = {1},
pages = {828--838},
doi = {10.1109/tce.2024.3518755}
}
EUVIMED