结合DCT和VQ编码的可恢复脆弱水印算法

Restorable Fragile Watermarking Algorithm Combined DCT with VQ Encoding

由于编码效率和安全性不够理想,现有可恢复脆弱水印算法的篡改恢复性能有待进一步提高,结合DCT(Discrete Cosine Transform)和VQ(Vector Quantization)编码提出了一种编码方式可变的变容量可恢复水印算法。为了提高编码效率,根据8×8图像块的复杂程度,分别采用DCT或VQ编码生成变容量重要恢复水印信息。对4×4的高频分量块,利用VQ压缩生成8比特细节水印。将图像块的重要水印分成三部分及其细节水印基于密钥分别随机地嵌入在其他图像块中,以抵抗拼贴攻击。篡改检测时,综合四部分水印的不一致标识,判定图像块的真实性。为了提高恢复性能,利用提取的重要、细节水印信息和图像修复方法恢复篡改图像块。仿真结果表明,算法具有较好的编码效率和安全性,并且篡改恢复性能优于现有文献。

Due to the dissatisfactory coding efficiency and security,the tamper recovery performance of existing fragile watermarking schemes can be further improved.Combined DCT and VQ encoding,an alterable-capacity watermarking scheme with restoration capability is proposed,in which the encoding method can be alterable.To improve the coding efficiency,DCT or VQ encoding method is selected to generate the significant recovery watermark with alterable-capacity according to the complexity of each 8 × 8 image block.For each 4 × 4 high-frequency component block,the detail watermark with eight bits is generated by VQ compression.Then,based on the user key,the significant watermark which is divided into three parts and the detail watermark of the block are embedded in other image blocks to resist collage attack.In the tamper detection,the proposed scheme synthesizes the inconsistent marks corresponding to four parts of watermark to judge the authenticity of blocks.To improve the tamper recovery performance,the extracted significant,detail watermark and image inpainting method are used to recover the tampered image blocks.Simulation results show that the proposed scheme can not only achieve better coding efficiency and security,but also improve the restoration capability.