Cao, Jie https://orcid.org/0000-0001-6899-8637
Cai, Song
Han, Daolong
Dong, Chao
Xu, Jianfeng
Lu, Guihua
Article History
Received: 11 June 2025
Accepted: 5 October 2025
First Online: 4 November 2025
Declarations
:
: The authors declare that there are no competing interests, financial or non-financial, that could be perceived as influencing the work reported in this paper.
: This study did not involve human participants, animal experiments, or sensitive biological materials requiring ethical approval. All public datasets utilized in this study–including ICDAR2015, CIFAR-10, and SVHN–have pre-completed anonymization processing (e.g., ICDAR2015 removes all personally identifiable information such as user-specific scene metadata, ensuring no individual privacy is involved). Additionally, the use of these datasets strictly complies with their official license terms: ICDAR2015 adheres to the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license, CIFAR-10 and SVHN follow the MIT License, and no commercial or unauthorized derivative use is conducted. The above practices fully meet the ethical standards for public dataset utilization in computer vision research.