在JPEG XR 影像上做資訊隱藏之新研究

Abstract

With the convenience of the Internet, information exchanges through the Internet become very frequent nowadays. The JPEG XR image is a new image file format which supports high compression ratios and becomes more and more popular in information communication. In this study, three methods for data hiding applications via JPEG XR images are proposed, including covert communication, image authentication, and copyright protection. For covert communication, a data hiding method via JPEG XR images for use in covert communication is proposed, which encodes selected quantization parameters of highpass frequency bands to embed secret messages. In addition, a technique to generate a stego-image with the original cover image file size being kept for reducing the possibility of arousing notice from attackers is also proposed. For image authentication, a block-based authentication method based on embedding DC_LP coefficients as authentication signals into the highpass coefficients of the image blocks is proposed, which can be employed to verify the integrity and fidelity of JPEG XR images by comparing the extracted authentication signals with those computed from the current image content. For copyright protection, a removable visible watermarking method based on modifying the DC_LP coefficients in JPEG XR images by certain number division operations is proposed. Specifically, the remainders of dividing the DC_LP coefficients by a pre-selected integer number are embedded into the highpass coefficients of image blocks as watermark signals which can be losslessly removed if necessary. Measures to enhance the security of each of the proposed methods are also suggested. Good experimental results show the feasibility of the proposed methods.