Content-Based Secure Image Retrieval in an Untrusted Third Party Environment

EasyChair Preprint 9037

Abstract

In this digital world, where availability of the image-generating tools is quite common and owing to the rapid growth of internet knowledge; people use to exchange massive volume of images every day which results in creating large image repositories. So, retrieving appropriate image available on these repositories is one of the vital tasks. This problem leads to evolving content-based image retrieval (CBIR). As the generation of image increases, people start transferring these images to a remote third-party server, but these images may have personal information. This leads to adding privacy concerns toward the system as transferring personal data to some other place might be a cause of leakage of information or transfer to an unauthorized person. So, to keep this in mind, sensitive images like medical and personal images require encryption before being a contracted out for the privacy-preserving resolutions. In this work, we have deployed ACM for image encryption as well as Asymmetric Scalar Product Preserving Encryption (ASPE) for feature vector encryption and similarity matching. We have demonstrated our results based on various benchmark databases.

Keyphrases: ACM, Image encryption, Local Binary Pattern, feature extraction, image retrieval

@booklet{EasyChair:9037,
  author    = {Sandeep Singh Sengar and Sumit Kumar},
  title     = {Content-Based Secure Image Retrieval in an Untrusted Third Party Environment},
  howpublished = {EasyChair Preprint 9037},
  year      = {EasyChair, 2022}}