A novel intelligent robust watermarking scheme for protecting information in document images

In most of the document images, there are segments containing variations of generatable and non-generatable information content. Hence, variable amount of watermark can be applied to these segments of the document image. This could reduce the time required for embedding the watermark. In this paper, a novel intelligent robust watermarking scheme is proposed that intelligently embeds the watermark into the segments of a document image. Thus varying amount of watermark can be embedded into different segments of document image. The proposed watermarking scheme consists of three phases. In Phase-1, the segments of the document image are intelligently identified into different categories as completely generatable, variations of generatable and non-generatable information content and completely non-generatable. This classification is performed by extracting Grey Level Co-occurrence Matrix (GLCM) features of each segment and computing entropy values. In Phase-2, embedding of the watermark is performed using integer wavelet based embedding mechanism. The watermark to be embedded is a binary logo which is compressed using Binary Block coding technique with blocks of uniform size. In Phase-3, the bitstream is extracted from each segment of the watermarked document image and decoded to get the binary watermark. The extracted and embedded watermarks are compared and authentication is verified using majority voting technique [13]. The experimental results of the proposed watermarking scheme exhibit higher success rate in identification of the type of segment. The performance of the proposed approach is evaluated in terms of Peak Signal to Noise Ratio (PSNR) and time taken for embedding watermarks. The comparative analysis reveals that the proposed method outperforms than the existing methods [18,25].