Access Control with Encrypted Feature Maps for Object Detection Models
Summary :
In this paper, we propose an access control method with a secret key for object detection models for the first time so that unauthorized users without a secret key cannot benefit from the performance of trained models. The method enables us not only to provide a high detection performance to authorized users but to also degrade the performance for unauthorized users. The use of transformed images was proposed for the access control of image classification models, but these images cannot be used for object detection models due to performance degradation. Accordingly, in this paper, selected feature maps are encrypted with a secret key for training and testing models, instead of input images. In an experiment, the protected models allowed authorized users to obtain almost the same performance as that of non-protected models but also with robustness against unauthorized access without a key.
- Publication
- IEICE TRANSACTIONS on Information Vol.E106-D No.1 pp.12-21
- Publication Date
- 2023/01/01
- Publicized
- 2022/11/02
- Online ISSN
- 1745-1361
- DOI
- 10.1587/transinf.2022MUP0002
- Type of Manuscript
- Special Section PAPER (Special Section on Enriched Multimedia--Advanced Safety, Security and Convenience--)
- Category
Authors
Teru NAGAMORI
Tokyo Metropolitan University
Hiroki ITO
Tokyo Metropolitan University
AprilPyone MAUNGMAUNG
Tokyo Metropolitan University
Hitoshi KIYA
Tokyo Metropolitan University
Keyword
Latest Issue
Copyrights notice of machine-translated contents
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Cite this
Copy
Teru NAGAMORI, Hiroki ITO, AprilPyone MAUNGMAUNG, Hitoshi KIYA, "Access Control with Encrypted Feature Maps for Object Detection Models" in IEICE TRANSACTIONS on Information,
vol. E106-D, no. 1, pp. 12-21, January 2023, doi: 10.1587/transinf.2022MUP0002.
Abstract: In this paper, we propose an access control method with a secret key for object detection models for the first time so that unauthorized users without a secret key cannot benefit from the performance of trained models. The method enables us not only to provide a high detection performance to authorized users but to also degrade the performance for unauthorized users. The use of transformed images was proposed for the access control of image classification models, but these images cannot be used for object detection models due to performance degradation. Accordingly, in this paper, selected feature maps are encrypted with a secret key for training and testing models, instead of input images. In an experiment, the protected models allowed authorized users to obtain almost the same performance as that of non-protected models but also with robustness against unauthorized access without a key.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2022MUP0002/_p
Copy
@ARTICLE{e106-d_1_12,
author={Teru NAGAMORI, Hiroki ITO, AprilPyone MAUNGMAUNG, Hitoshi KIYA, },
journal={IEICE TRANSACTIONS on Information},
title={Access Control with Encrypted Feature Maps for Object Detection Models},
year={2023},
volume={E106-D},
number={1},
pages={12-21},
abstract={In this paper, we propose an access control method with a secret key for object detection models for the first time so that unauthorized users without a secret key cannot benefit from the performance of trained models. The method enables us not only to provide a high detection performance to authorized users but to also degrade the performance for unauthorized users. The use of transformed images was proposed for the access control of image classification models, but these images cannot be used for object detection models due to performance degradation. Accordingly, in this paper, selected feature maps are encrypted with a secret key for training and testing models, instead of input images. In an experiment, the protected models allowed authorized users to obtain almost the same performance as that of non-protected models but also with robustness against unauthorized access without a key.},
keywords={},
doi={10.1587/transinf.2022MUP0002},
ISSN={1745-1361},
month={January},}
Copy
TY - JOUR
TI - Access Control with Encrypted Feature Maps for Object Detection Models
T2 - IEICE TRANSACTIONS on Information
SP - 12
EP - 21
AU - Teru NAGAMORI
AU - Hiroki ITO
AU - AprilPyone MAUNGMAUNG
AU - Hitoshi KIYA
PY - 2023
DO - 10.1587/transinf.2022MUP0002
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E106-D
IS - 1
JA - IEICE TRANSACTIONS on Information
Y1 - January 2023
AB - In this paper, we propose an access control method with a secret key for object detection models for the first time so that unauthorized users without a secret key cannot benefit from the performance of trained models. The method enables us not only to provide a high detection performance to authorized users but to also degrade the performance for unauthorized users. The use of transformed images was proposed for the access control of image classification models, but these images cannot be used for object detection models due to performance degradation. Accordingly, in this paper, selected feature maps are encrypted with a secret key for training and testing models, instead of input images. In an experiment, the protected models allowed authorized users to obtain almost the same performance as that of non-protected models but also with robustness against unauthorized access without a key.
ER -
English
