Privacy preserving LBP based feature extraction on encrypted images
Sayyada Fahmeeda Sultana, D C Shubhangi · 2017
“An Image is worth a thousand words.” To provide security images always encrypted and stored in the cloud. Along with it, there is also a continuous need to outsource image computation with high complexity to the cloud for its economic computing resources and on-demand ubiquitous access. Feature extraction and representation on encrypted images is a crucial step for multimedia processing on the cloud. Extraction of ideal features from encrypted images without revealing the intrinsic content of the images is privacy preservation. In this paper, we propose an efficient scheme PP-LBP (Privacy preserving - Local Binary Patterns) that retrieves LBP based image features from encrypted images. In the Proposed approach, the encryption algorithm is applied on MSB (Most Significant Bit) plane of Image. All the operations are performed on encrypted images without revealing any information to cloud service providers. The method generates exactly same LBP based feature between encrypted and unencrypted images without any overhead of communication between cloud service provider and client. All the Computational overheads complexities are on cloud not to client.