Abstract
With the tremendous growth and usage of digital images nowadays, the integrity and authenticity of digital content is becoming increasingly important and of major concern to many government and commercial sectors. In the past decade or so, digital watermarking has attracted much attention and offers some real solutions in protecting the copyright and authenticating the digital images. Recently, image forensics techniques, based on a passive statistical analysis of the image data only, is an alternative approach to the active embedding of data associated with digital watermarking. In this thesis, the concept of digital watermarking is described, and a number of algorithms used for robust, fragile and semi-fragile image watermarking are reviewed in detail. We discuss the concept of image forensics briefly and conclude image forensics techniques into two main areas: image source detection and image manipulation detection. Four novel robust and semi-fragile transform based image watermarking related schemes are introduced. These include wavelet-based contourlet transform (WBCT) for both robust and semi-fragile watermarking, slant transform (SLT) for semi-fragile watermarking as well as applying the generalised Benford’s Law to estimate JPEG compression, then adjust the appropriate threshold for improving the semi-fragile watermarking technique. In this thesis, the proposed WBCT for robust watermarking is evaluated and compared with two other Discrete Wavelet Transform (DWT) based algorithms with results achieving high degree of robustness against most non-geometrical and geometrical attacks, while maintaining an excellent perceptual quality. For semi-fragile watermarking, the proposed SLT as a block-based algorithm achieves more accuracy for copy & paste attacks with non-malicious manipulations, such as additive Gaussian noise when compared with existing Discrete Cosine Transform (DCT) based and Pinned Sine Transform based schemes. While for the proposed WBCT method, good performance is achieved in localising the tampered regions, even when the image has been subjected to non-malicious manipulations such as JPEG/JPEG2000 compressions, Gaussian noise, Gaussian filtering, and contrast stretching. The average false negative rate is found to be approximately 1% while maintaining an average false positive rate below 6.5%. We also propose the use of generalised Benford’s Law model as an image forensics technique for semi-fragile watermarking. This model can improve the lower tampered detection rate caused by the predetermined threshold in semi-fragile watermarking. The threshold is typically fixed and cannot be easily adapted to different amounts of errors caused by unknown JPEG compression. Our proposed method can adaptively adjust the threshold for images based on the estimated Quality Factor (QF) by using the generalised Benford’s Law with overall average QF correct detection rate of approximately 99% when 5% of the pixels axe subjected to image content tampering, as well as compression using different QFs (ranging from 95 to 65).