Development of Efficient Watermarking Methods to Enhance Security

Abstract

The technological advancement has provided an opportunity to share the various multimedia content globally. However, these multimedia content like image, text, audio, and video can be easily manipulated and transferred among the public. For data integrity, digital watermarking is an efficient technique which protects multimedia contents by hiding a watermark. Watermarking technique offers ownership and authenticity by hiding a watermark in images videos. Several watermarking techniques have been developed in the literature that embed watermark into spatial domain or frequency domain to protect the multimedia contents. Generally, frequency domain techniques are often used due to its robustness. However, some limitations have been observed in various watermarking techniques such as single embedding factor, false positive problem, low embedding capacity, and others. Therefore, the primary goal of this work is to design and develop efficient watermarking methods to improve security, robustness, embedding capacity, and imperceptibility. This thesis divides the presented work into five folds. Firstly, a new modified gravitational search algorithm has been developed to balance the trade off among the watermarking parameters. Secondly, an optimal multiple embedding factor selection based watermarking technique is introduced. Third, an optimized watermarking technique is presented to overcome the problem of false positive error. The first two watermarking techniques are implemented for grayscale images and embed smaller sizes of watermark images. Therefore, the fourth part introduces a high embedding capacity based color image watermarking technique. Finally, this work presents a robust video watermarking technique for ownership verification newline