Wavelet Based Robust Watermarking Scheme

Abstract

Digital Watermarking is an established security modality with many application. Its performance is influenced by different intentional and unintentional attacks. This thesis focuses on removing these challenges by developing efficient embedding and extraction algorithm and enhance the security system against various attacks.
Digital watermarking is the process of embedding a watermark in a multimedia content like images, video, audio or any digital content for information hiding. With the help of digital watermarking technique, we can protect the multimedia content against different illegal actions. The watermark provides the authentication of the owner data. Here, the data hiding scheme hides the authorized logo with the original image. These techniques have been proposed for different applications like copyright protection, content authentication, ownership protection and tamper detection.
The motivation of this thesis is to increase the watermark strength and provide a better trade-off between imperceptibility and robustness of the watermarked image. For this purpose, we have discussed a robust watermarking wavelet technique called Lifting Wavelet Transform (LWT) in the frequency domain. Here low-pass sub-band is used for data hiding process. The coefficients of low-pass sub-band are randomly shuffled by using the secret key. By the help of block selection procedure, the low-pass sub-band is divided into small blocks. After that two maximum coefficients of each block quantizing to embedded binary watermark with the host image. During the extraction process, the reverse lifting technique is followed. As per the concerned security randomization of coefficients, blocks, secret key and lifting technique enhanced the security level of the system.