Tamper detection technique for document images using zero watermarking in wavelet domain

Existing tamper detection techniques for document images use the textual information of the image to create a zero watermark. These techniques cannot authenticate a document image if the image is tampered with in non-textual regions. In order to resolve this problem, a tamper detection technique for document images is proposed using the concept of zero watermarking. The input document image is transformed by applying lifting wavelet transform to get its subbands. These subbands are further partitioned into non-overlapping blocks of the same size. Features of each blocks are extracted to create a zero watermark. This watermark is combined with the key watermark in order to create a meaningful watermark. This meaningful watermark is stored by a trusted authority. On the receiver side, the meaningful watermark is compared to the watermark generated using tampered image to check the authenticity of the cover image. The tampered regions of the cover image can also be localized using this technique. The experimental results demonstrate the effectiveness of the technique in comparison to the existing authentication techniques.