基于值合并及混沌映射的JPEG图像精确认证方案  

JPEG image authentication scheme based on value combination technique and chaotic maps

作　　者：张杰[1] 刘如玉[1] 王国栋[1] 汪萍[1] 刘粉林[1] 

机构地区：[1]解放军信息工程大学信息工程学院,郑州450002

出　　处：《东南大学学报（自然科学版）》2007年第A01期114-118,共5页Journal of Southeast University：Natural Science Edition

基　　金：国家高技术研究发展计划(863计划)资助项目(2006AA01Z409);河南省科技攻关资助项目(0623021500)

摘　　要：为了提高JPEG图像认证算法的水印嵌入容量及安全性问题,提出一种新的JPEG图像精确认证方案.通过调整标准量化表进而调整DCT系数,以增加水印的嵌入容量,同时,对DCT系数及位置信息进行值合并运算并得到合并值,将其作为混沌系统的初值经过迭代后产生水印,并将水印嵌入调整后的JPEG系数中,最后通过熵编码形成含水印的JPEG图像.理论分析和实验结果表明:该方案可以有效折中图像质量和嵌入容量之间的矛盾,有较高的安全性,同时可以准确定位图像被篡改的区域.A novel authentication algorithm is proposed to improve the watermark embedding capacity and security for JPEG（joint picture experts group） image. The discrete cosine transform （DCT） coefficient is adjusted via the modification of standard quantization table to improve the capacity. Then the coefficient is combined with the position information to obtain the combined value. And the watermark is generated with the combined values as the initial state of chaotic system, then the watermark is embedded into the corresponding DCT coefficients. Furthermore, the watermarked JPEG image is produced by entropy encoding. Theoretical analysis and simulation results show that the proposed algorithm can effectively tradeoff the conflict of embedding capacity and image quality and locate tampered blocks accurately. Besides, it is more secure.

关 键 词：脆弱水印 量化攻击 值合并 篡改定位 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]