CAJ | 학술논문

根据压缩感知具有压缩短、解码复杂和鲁棒性强的特征，提出一种新的基于压缩感知的数字图像水印算法。该算法有效克服了信号重建时稀疏度未知的问题，增强了数字水印系统的抗提取防破坏能力。在嵌入水印时，将水印的高频系数用构造的伯努利矩阵进行观测后，嵌入于图像的低频系数中；在提取水印时，利用正则化自适应匹配追踪（RAMP）算法对高频系数进行恢复，最后与低频系数结合实现水印的重构。实验结果表明：该算法在重构精度和重构时间上要高于同类算法，具有很好的鲁棒性、透明性和安全性。
근거압축감지구유압축단、해마복잡화로봉성강적특정，제출일충신적기우압축감지적수자도상수인산법。해산법유효극복료신호중건시희소도미지적문제，증강료수자수인계통적항제취방파배능력。재감입수인시，장수인적고빈계수용구조적백노리구진진행관측후，감입우도상적저빈계수중；재제취수인시，이용정칙화자괄응필배추종（RAMP）산법대고빈계수진행회복，최후여저빈계수결합실현수인적중구。실험결과표명：해산법재중구정도화중구시간상요고우동류산법，구유흔호적로봉성、투명성화안전성。
Since compressive perception has the characteristics of short compression,decoding complexity and strong ro-bustness,a new digital image watermark algorithm based on compressive perception is proposed. The algorithm overcame the shortcoming that the sparsity is unknown in signal reconstruction,and enhanced the function to resist illegal extraction and pre-vent damage of digital watermark system. When watermark is embedded,the high-frequency coefficients of the watermark are measured by the constructed Bernoulli matrix,and then the watermark is embedded in the low-frequency coefficients of the image. As for the watermark extraction,the regularized adaptive matching pursuit(RAMP)algorithm is used to recover the high-frequency coefficients,and make them combined with the preserved low-frequency coefficients to obtain the reconstructed water-mark. Compared with the original compressed sensing algorithm,The experimental results demonstrate that the proposed algo-rithm is better than the original compressive perception algorithm in reconstruction accuracy and time consumption,and has good robustness,high transparency and security.