Enhancement of most significant bit (MSB) algorithm using discrete cosine transform (DCT) in non-blind watermarking / Halimah Tun Abdullah

• Main Author: Abdullah, Halimah Tun
• Format: Thesis
• Language: English
• Published: 2014
• Subjects: Electronic Computers. Computer Science; Integrated software; Algorithms
• Online Access: https://ir.uitm.edu.my/id/eprint/65339/2/65339.pdf; https://ir.uitm.edu.my/id/eprint/65339/

Digital watermarking is used to protect the intellectual property from malicious attack. This is because everyday people easily take the images or download the images without permission from ownership. Digital watermarking can hide data, such as in image, video and audio. In this project, the image had been chosen as the secrecy of the data. This is because many attackers attempt to retrieve the secret data from an image than audio and video. The algorithm and techniques that implemented in this project is Most Significant Bit (MSB) algorithm and Discrete Cosine Transform (OCT) as a technique for embedding process. According to how watermark detected and extracted, this project use non-blind watermarking mean need an original image during extraction process. The non-blind watermarking is more secure and can endure in a variety of attacks such as compression, rotation, cropping, filtering and others. In this project, an image from different types was used during the embedding process. Then, the result had being compared between the different types of images. The types of images are Joint Photographic Experts Group (JPEG), Bitmap (BMP), and Portable Network Graphics (PNG). This project had been done in order to test the quality watermarked image using Peak Signal to Noise Ratio (PSNR). Findings show bitmap image is a high quality watermarked image from the other types of images since its value is more than 30 decibels (dB). The robustness of the watermarked image will test by attacking the watermarked image as the enhancement of this project.