Securing SIFT: Privacy-Preserving Outsourcing Computation of Feature Extractions Over Encrypted Image Data

Area of Science:

• Cloud Computing Security
• Computer Vision
• Cryptography

Background:

• Cloud computing offers cost savings and flexibility for data storage and processing.
• Outsourcing multimedia data to the cloud risks exposing private information.
• Privacy-preserving computations are crucial for secure cloud-based data analysis.

Purpose of the Study:

• To develop an effective and practical privacy-preserving computation outsourcing protocol for Scale-Invariant Feature Transform (SIFT).
• To address the efficiency, security, and practicality limitations of existing solutions.
• To preserve the distinctiveness and robustness characteristics of the original SIFT algorithm.

Main Methods:

• Randomly splitting original image data.
• Designing novel efficient protocols for secure multiplication and comparison.
• Distributing feature extraction computations across two independent cloud servers.

Main Results:

• The proposed scheme achieves efficiency and security simultaneously.
• The protocol preserves key SIFT characteristics like rotation and scale invariance.
• Evaluations demonstrate practical security and superior performance compared to state-of-the-art methods.

Conclusions:

• The developed protocol offers a secure and efficient solution for privacy-preserving SIFT computation on encrypted cloud data.
• It maintains the robustness and distinctiveness of SIFT features, comparable to the original algorithm.
• This work enables secure outsourcing of computationally intensive image analysis tasks to the cloud.