A Data Hiding Method for Few-Color Images

Gang Pan,Yijun Wu,Zhaohui Wu
DOI: https://doi.org/10.1007/3-540-45600-7_30
2001-01-01
Abstract:Binary images have only two colors, which makes the embedding of invisible data difficult. In this paper, we propose a new data hiding method that can hide a moderate amount of data in a host binary image, such as binary cartoon images, scanned texts, signatures, without introducing noticeable artifacts. The proposed method employs subblock pattern classification to maintain visualization effect and mechanics of multilevel supblock to improve the capacity. Extracting of the hidden data does not require the knowledge of the original image. The experiments demonstrate that the proposed method can provide excellent perceptual quality of the marked image. The potential applications include invisible annotation, alteration detection and covert communication.
What problem does this paper attempt to address?