An efficient privacy preserving medical image retrieval using ROI enabled searchable encryption

Abstract

In this paper, we propose a new idea of searchable medical image encryption method to provide secrecy or authentication when a database is stashed away on a host maintained by a third-party server. This project addresses the trouble of enabling content based image retrieval throughout encrypted medical image dataset. Medical image database is encrypted by the owner of the content before transmitting or before stacked away on to the server. These encrypted stored data along with patient information are made accessible for authorized users only. The authorized user can pose the query on the encrypted data base and able to retrieve the most relevant cluster of images from the database maintained by the third part server. We compare our proposed privacy preserving data mining technique with conventional technique in terms of precision and recall. The insights obtained through this research and comparisons will help to design real time algorithms appropriate for privacy aware third party server systems.