A privacy-preserving, on-board satellite image classification technique incorporating homomorphic encryption and transfer learning

December 10, 2025Frontiers in Remote SensingOpen Access

View Full Paper

Ask AI

Bookmark

Share

Key Points

The aim is to enhance on-board satellite image classification while ensuring data privacy through encryption.
Utilized homomorphic encryption for secure storage of satellite images in cloud storage.
Implemented transfer learning models including YOLO, ResNet, and vision transformer for classification.
Training was conducted separately with decrypted images to develop the classification module.
The best-trained model was encrypted and limited to authorized devices.
The vision transformer classifier achieved the highest accuracy of approximately 99.65%.
The approach successfully maintains image privacy while enabling effective on-board classification.

We only have this paper's title and publication details.

Answers are grounded in this paper and cite their sources.

Hybrid approach improves privacy and accuracy in satellite image classification with trained models, indicating effective cloud storage use.

No takes yet. Share an insight, caveat, or question.

Roy et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

Share