A HYBRID BRIEF-SVD WATERMARKING TECHNIQUES FOR DATA PROTECTION IN CLOUD COMPUTING

Abstract

In the recent world with developing technologies, information is collected, stored digitally, and transmitted. While transmitting the information digitally, security and authenticity remain the main issue. Watermarking is the technology that ensures the authenticity and security of multimedia images. Digital watermarking techniques played an essential role in protecting and authenticating the copyright of multimedia content. There are several digital watermarking techniques used for image and data protection in cloud computing. However, because of the poor image quality, the watermarking techniques currently in use are inefficient, which presents a security challenge. The crucial requirements for designing an efficient watermarking scheme are robustness, imperceptibility, capacity, and security. To meet all of these requirements simultaneously is nearly impossible. Therefore, this research study developed a hybridized watermarking technique that is Binary Robust Independent Elementary Features (BRIEF) and Singular Value Decomposition (SVD), in order to enhance robustness, imperceptibility, and security. The proposed hybridized scheme was used to perform experiment using image data obtained from Kaggle Repository. The performance of the proposed hybrid technique was subjected to an evaluation using four metrics namely; Peak Signal to Noise Ratio (PSNR), Signal to Noise Ratio (SNR), Mean Square Error (MSE), and Means Absolute Error (MAE). Thus, the following results were obtained: PSNR value of 34.29dB, SNR of 11.44dB, MSE 2.8 MAE 6.88 respectively, making it the best performing technique compared to that of the existing techniques which were only based on PSNR with DCT value of 30dB and FRFT, DCT value of 33.518dB. BRIEF-SVD performed efficiently well in terms of PSNR, SNR, MSE, and MAE. This study concludes that the hybridization technique and use of feature descriptors are more robust and secure for image multimedia contents.