Self-embedding Watermarking Algorithm For Tamper Detection And Recovery

With the development of network technology, it is easy for offenders to tamper, edit and spread digital images through internet. It is difficult for people to identify the authenticity and integrity of received images because the offenders may tamper these digital images without leaving traces. If the tampered images are related to evidences or interests, the individual or group will suffer great economic loss or damage their personal visualize because tampered images distort the fact. How to identify the authenticity of the digital image and restore the truth through the tamper location and tamper recovery, that’s how to protect authenticity and integrity of the digital media content in the network environment has become a serious problem.The main contents of this thesis include:(1) A method for selecting watermark’s embedding position is proposed. Theoretical analysis shows that this method can increase the image scrambling effect, increase the secret key search space and improve the security of the secret key.(2) A cross-certification dual watermarking algorithm based on spatial domain is proposed. Firstly, the image is divided into blocks with the size of 2×2.Secondly, extract the features of each two image blocks and encrypt the features, then embed the encrypted features into mapping blocks. Lastly, extract the mapping block’s feature and embed the authentication watermark into its own. In the detection of tampered image, authentication watermarking not only certificates the elements within the its own block, but also certificates the corresponding block and elements within it. Theoretical analysis and experimental results show that the algorithm has good tamper localization and recovery capability. This algorithm is ideal for small area of shear attack, and it is much better than Zhao and Tong algorithms in tamper detection and tamper recovery when watered digital images suffer from collage attack and blind attack.(3) A hierarchical watermarking algorithm based on the combination of the spatial and frequency domain is proposed. This algorithm improves the cross certification dual watermarking algorithm based on the spatial domain. The proposed algorithm contains two levels in embedding watermark. In the 1-level:the authentication watermark consists of parity check code and XOR check code of 2×2 block and parity check code of its’ mapping block which is operated with slipping function; the recovery watermark consists of the gray value. Embed the watermark into mapping block. In the 2-level:the image processed in 1-level is divided into blocks with size of 8×8 extract frequency information and embed it into mapping block. The experimental results show that this algorithm not only has good effect in tamper detection and temper recovery when image suffers from blind attack or collage attack but also have good effect in temper recovery when the image is suffered from shear attack. This algorithm has an ideal recovery effect when the shear area is less than 56% of the total area. In terms of resisting various attacks, this algorithm is ideal for tampering localization and tampering. The algorithm is better than Tong’s algorithm and Zhao’s algorithm whose algorithms only have good ability in anti shear.