Image security attribute annotation technology based on CAN and WordNet

XiaoChang Wang; Na Wang; XueHui Du

doi:10.1117/12.2628042

18 February 2022 Image security attribute annotation technology based on CAN and WordNet

XiaoChang Wang, Na Wang, XueHui Du

Proceedings Volume 12162, International Conference on High Performance Computing and Communication (HPCCE 2021); 1216216 (2022) https://doi.org/10.1117/12.2628042
Event: 2021 International Conference on High Performance Computing and Communication, 2021, Guangzhou, China

Abstract

This paper propose an image security attribute annotation technology based on CAN and WordNet to realize the flexible, controllable and automatic annotation of dynamic emerging image security attributes, and providing the basic support of security attributes for dynamic and accurate access control based on ABAC (Attribute Based Access Control). Firstly, the transfer-learning model CAN (Contractual Adaptation Network) with CBAM (Convolutional Block Attention Module) attention mechanism used to extract the entity labels of image data. Secondly, based on the extracted entity label set, propose a security attribute tree construction method based on WordNet. This method can construct a security attribute tree with continuous semantic generalization and coarsening granularity layer by layer, to ensure that the scale of security attribute space is controlled and meet the access control requirements of different granularity. The experiment results show that the annotation accuracy of this technology can reach an average of more than 91% under different control granularity.

Citation Download Citation

XiaoChang Wang, Na Wang, and XueHui Du "Image security attribute annotation technology based on CAN and WordNet", Proc. SPIE 12162, International Conference on High Performance Computing and Communication (HPCCE 2021), 1216216 (18 February 2022); https://doi.org/10.1117/12.2628042

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available