Information Hiding Schemes by Using Difference Characteristics and Neighboring Pixel Prediction

• Main Authors: Wei-Shin Lin, 林唯歆
• Other Authors: Tzu-Chuen Lu
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/72971163047758369998

碩士 === 朝陽科技大學 === 資訊管理系碩士班 === 100 === As methods of information transmission are increasingly diversified, it becomes a critical concern to transmit information securely in open network environments, as well as to protect information from being easily retrieved even when it is stolen by unlawful third parties. Researchers proposed information hiding techniques to solve the problems related to transmitting private messages, in which secret messages are hidden in cover media to escape third party detection, thus achieving secure message exchange. Many information hiding schemes have been proposed. One of the famous methods is Modulus Function which proposed by Zhang and Wang. The method is effective and simple. However, the hiding capacity is limited. Hence, we attempt to improve the Modulus Function method by using the difference characteristics. The proposed method embeds information in the difference values between the maximum value and the remaining pixels in a 2 × 2 block. Experiment results show that the method can effectively increase the embedding capacity of the Modulus Function method. After that Lin proposed an information hiding scheme with minimal image distortion of embedded information. Lin’s scheme could effectively reduce pixel changes resulting from slight movement between pixels. However, their scheme also has hiding capacity problem. Therefore, we propose a high capacity method to improve the method of Lin. The proposed method combined the side matching method of Chang and Tseng with the improved minimal image distortion methods of Lin. It is based on the method of Lin. image processing was performed on three levels. First, the method of Lin was used to embed secret data into seed blocks. Next, the neighboring pixel anticipation method was applied to other pixels. Finally if the image has more pixels not embedded with secret data, the method of Lin was used to embed data into the image. Experimental results show that the proposed scheme outperforms the method of Lin.