Privacy-Preserving Semantic Segmentation Using Vision Transformer
Open Access
- 30 August 2022
- journal article
- research article
- Published by MDPI AG in Journal of Imaging
- Vol. 8 (9) , 233
Abstract
In this paper, we propose a privacy-preserving semantic segmentation method that uses encrypted images and models with the vision transformer (ViT), called the segmentation transformer (SETR). The combined use of encrypted images and SETR allows us not only to apply images without sensitive visual information to SETR as query images but to also maintain the same accuracy as that of using plain images. Previously, privacy-preserving methods with encrypted images for deep neural networks have focused on image classification tasks. In addition, the conventional methods result in a lower accuracy than models trained with plain images due to the influence of image encryption. To overcome these issues, a novel method for privacy-preserving semantic segmentation is proposed by using an embedding that the ViT structure has for the first time. In experiments, the proposed privacy-preserving semantic segmentation was demonstrated to have the same accuracy as that of using plain images under the use of encrypted images.Keywords
Funding Information
- JSPS KAKENHI (JP21H01327)
- Support Center for Advanced Telecommunications Technology Research, Foundation (SCAT) (JP21H01327)
This publication has 29 references indexed in Scilit:
- Privacy-Preserving Deep Learning via Weight TransmissionIEEE Transactions on Information Forensics and Security, 2019
- Semantic Understanding of Scenes Through the ADE20K DatasetInternational Journal of Computer Vision, 2018
- Encryption-Then-Compression Systems Using Grayscale-Based Image Encryption for JPEG ImagesIEEE Transactions on Information Forensics and Security, 2018
- An Efficient Convolution Core Architecture for Privacy-Preserving Deep Learning2018
- On the Security of Block Scrambling-Based EtC Systems against Extended Jigsaw Puzzle Solver AttacksIEICE Transactions on Information and Systems, 2018
- On the security of block scrambling-based ETC systems against jigsaw puzzle solver attacks2017
- The Cityscapes Dataset for Semantic Urban Scene Understanding2016
- Privacy-Preserving Logistic Regression with Distributed Data Sources via Homomorphic EncryptionIEICE Transactions on Information and Systems, 2016
- Unitary Transform-Based Template Protection and Its Application to l2-norm Minimization ProblemsIEICE Transactions on Information and Systems, 2016
- Privacy-preserving deep learning2015