Abstract: In Content-Based Image Retrieval (CBIR) the indexing of the image and the representation is done using visual contents of an image like color, shape, texture and spatial layout. Cloud storage is a popular choice to handle a large size data because its cost is much lower than hardware upgrade and infrastructure reorganization. Nowadays, image-based data is getting important in many applications such as face identification, disease detection and object recognition and it usually needs more storage than text-based data. However, the images usually contain personal and confidential information and hence directly outsourcing the image dataset to the cloud will arouse the privacy issue. To protect the sensitive information in images, it is necessary to encrypt images before being uploaded to the cloud. After storing the encrypted image data in the cloud, the users will ask the cloud server for the search of the target image data. However, current content-based image retrieval technologies are usually useless for searching encrypted image data, and hence this system tries to solve the problem of searching encrypted images in a large database.
Keywords:k-Nearest Neighbour( kNN ) , Homomorphic Encryption( HE ) , Asymmetric Scalar Product Preserving Encryption( ASPE ) , Content Based Image Retrieval( CBIR ) , Text Based Image Retrieval( TBIR )
| DOI: 10.17148/IJARCCE.2021.10735