Semi-Fragile Image Watermarking, Authentication and Localization Techniques for Law Enforcement Applications

Abstract

With the tremendous growth and use of digital cameras and video devices, the need to verify the collected digital content for law enforcement applications such as crime scene investigations and traffic violations, becomes paramount if they are to be used as evidence in courts. Semi-fragile watermarking has become increasingly important within the past few years as it can be used to verify the content of images by accurately localising the tampered area and tolerating some non-malicious manipulations. There have been a number of different transforms used for semi-fragile image watermarking. In this chapter, we present two novel transforms for semi-fragile watermarking, using the Slant transform (SLT) as a block-based algorithm and the wavelet-based contourlet transform (WBCT) as a non-block based algorithm. The proposed SLT is compared with existing DCT and PST semi-fragile watermarking schemes. Experimental results using standard test images and simulated law enforcement images indicate that the SLT is more accurate for copy and paste attacks with non-malicious manipulations, such as additive Gaussian noise. For the proposed WBCT method, watermarking embedding is performed by modulating the parent-children relationship in the contourlet domain. Again, experimental results using the same test images have demonstrated that our proposed WBCT method achieves good performances 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 miss detection rate is found to be approximately 1% while maintaining an average false alarm rate below 6.5%.