ABSTRACT
The rapid growth of digital image repositories necessitates efficient Content-Based Image Retrieval (CBIR) systems that rely on computer vision technology to store and search massive amounts of digital information. Since the capabilities of digital storage increases it is necessary to protect digital information from unauthorized access and it's an important concern. Although new encryption techniques have evolved, numerous approaches proposed for cryptographic algorithms have proven to be insecure, and as such, sensitive information remains at risk of exposure. Choosing the best-suited encryption algorithm is crucial for data security, yet it is a time-consuming process. This paper introduces a Robust Data Hiding (RDH) method, along with a Triple DES block-based transformation algorithm, for image content security protection. The developed scheme supports secure image storage and communication, as well as content-based image retrieval and image convolution operations directly on the encrypted images. This ability is supported by cloud computing platforms, allowing distributed computational work. Security analysis proves that the given scheme provides good protection in both statistical and computational points of view. Though conventional cryptographic techniques provide greater data confidentiality, the proposed technique is designed for common users with additional operations like image search and convolution. Experimental observations affirm the effectiveness of encryption, retrieval, and convolution stages at an acceptable cost of storage. In this research, a workable implementation for offline image search on desktop computers and a foundation for subsequent CBIR systems with comparable uses is presented. As the growth of multimedia keeps on increasing in today's digital age, there has become an important task for secure and effective image retrieval and authentication systems.
