A Steganography Approach for Printed Image Based on Image Complexity and Template Matching

Abstract

After a digital image is discretized, printed and rescanned, it is usually filtered, rotated, scaled, cropped, con- trast-and-luminance adjusted, as well as distorted by noises. Robustness and invisibility against distortions caused by im- age processing and printing are the essential properties for image watermarking to be applicable in real-world applications. The existing watermark methods used in printed images are less than satisfactory in imperceptibility and robustness. In this paper, the properties of the discretized, printed image and rescanned image in both the spatial and frequency domains are analyzed. Based on these properties, image complexity and template matching theories, this paper proposes a ste- ganography approach for data-hiding in printed image. First, the message to be hidden is coded by using BCH to facilitate the extraction of hidden information. Secondly, based on appropriate image block partition, the local image complexity of the block is calculated to decide whether to embed watermark in this block. Watermark is embedded on the mid-frequency spectrum in DFT domain. Thirdly, two structural template lines are added in the polar coordinate system for detecting any combination of the geometrical distortions. Then, the hidden data can be extracted without using the original image. The experimental results indicate that the proposed method is robust against standard image manipulation operations, basic at- tacks and image publishing system. And the robustness and the invisibility of watermark are also satisfactory in printed images.