Encrypted Remote Sensing Image Retrieval With Authenticity Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing remote sensing image retrieval systems in cloud environments face challenges in achieving a balance between retrieval efficiency, accuracy, security, and authenticity due to issues with feature extraction disparities among multiple owners, high-dimensional feature storage, insufficient security measures, and tampering by semi-honest cloud servers.
Innovation Solution
A method involving encrypted searchable indexes, Merkle trees, and watermark embedding to ensure secure and efficient retrieval, using CNN models for feature extraction, spectral hashing for dimensionality reduction, and ASPE encryption to protect privacy, with a certification authority verifying authenticity and legality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If remote sensing images are encrypted before outsourcing to cloud server, then image confidentiality is improved, but retrieval accuracy and efficiency deteriorate
Solution Approach 1:
The patent applies asymmetric scalar-product-preserving encryption (ASPE) to encrypt image features and query vectors before cloud processing. This preliminary encryption action allows the cloud server to perform similarity computations on encrypted data, ensuring confidentiality is maintained throughout the retrieval process while still enabling accurate matching through the mathematical properties of ASPE that preserve scalar products in encrypted space.
2Measurement precision
If high-dimensional features are extracted through deep learning, then feature extraction accuracy is improved, but storage and computational resources increase
Solution Approach 1:
The patent extracts only the essential feature vectors from deep learning models for encryption and storage, rather than storing complete high-dimensional feature sets. This extraction approach maintains the discriminative power needed for accurate retrieval while significantly reducing the storage burden on the cloud server and improving processing efficiency.
3Reliability
If watermarking technology is applied to trace data distribution, then data distribution tracking is improved, but image fidelity and retrieval flexibility deteriorate
Solution Approach 1:
The patent embeds watermarks in specific local regions of the encrypted images rather than applying global watermarking. This localized approach allows the watermark to be embedded in less critical areas, minimizing the impact on overall image fidelity while still enabling effective tracking of data distribution and user authentication.
4Adaptability or versatility
If cloud servers operate under semi-honest model, then system flexibility is improved, but data authenticity and security deteriorate
Solution Approach 1:
The patent introduces a certification authority as an intermediary that issues digital certificates to verify the authenticity of image data returned by the cloud server. This intermediary mechanism allows the system to maintain flexibility in cloud operations while ensuring data authenticity through cryptographic verification, mitigating the risks of semi-honest cloud behavior.
5Adaptability or versatility
If multiple data owners are involved, then data diversity is improved, but feature extraction consistency and training scalability deteriorate
Solution Approach 1:
The patent employs a universal ASPE encryption scheme that can process feature vectors from multiple different data owners and deep learning models consistently. This universal approach allows the system to accommodate diverse data sources and extraction methods while maintaining consistent encryption and comparison operations, enabling scalable multi-owner collaboration without sacrificing feature extraction consistency.
Data Source
AI summary
The disclosure discloses a remote sensing image security retrieval method and a remote sensing image security retrieval system in cloud environment. According to the present disclosure, a CNN model is applied to extract features, then dimension reduction is performed according to spectral hashing with spectral rotation, and clustering is performed through K-means++. A secure encryption searchable index is generated from the correction tuple of Asymmetric Scalar-Product-Preserving Encryption (ASPE). Image encryption is performed after creating watermark embedding space through prediction error marking and pixel rearrangement. The copyright information and the user information are flexibly embedded and extracted in the process for copyright authentication and distribution data tracking. In addition, the authenticity of the search result is verified by constructing the Merkle tree.