Related Experiment Video
Updated: Jun 9, 2025
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
475
A dynamic authorizable ciphertext image retrieval algorithm based on security neural network inference
Xin-Yu Zhang1, Jing-Wei Hong2,3
1School of Statistics and Applied Mathematics, Anhui University of Finance and Economics, Beng' bu, China.
Plos One
|October 23, 2024
Summary
This study introduces a secure image retrieval system using encrypted data and neural networks. It enhances privacy and allows flexible, authorized access to images without decryption.
Area of Science:
- Computer Science
- Information Security
- Artificial Intelligence
Background:
- Traditional image retrieval systems often compromise data privacy.
- Secure image retrieval is crucial for cloud computing environments.
- Existing methods may lack flexibility and dynamic authorization.
Purpose of the Study:
- To propose a dynamic, authorizable ciphertext image retrieval scheme.
- To enhance image retrieval security while preserving data privacy.
- To enable efficient and flexible retrieval of authorized encrypted images.
Main Methods:
- Utilizing secure neural network inference for feature extraction on encrypted images.
- Implementing a dynamic, authenticatable ciphertext retrieval algorithm.
- Developing an index construction stage for encrypted image features.
Main Results:
- Guaranteed data image privacy throughout the upload-to-retrieval process.
- Ensured data availability and security for convenient image retrieval.
- Demonstrated a practical solution for cloud computing environments, despite potential efficiency limitations.
Conclusions:
- The proposed scheme effectively enhances image retrieval security and privacy.
- Dynamic authorization and secure inference provide flexibility and robust protection.
- The solution meets practical needs for secure cloud-based image retrieval.
