The Vector Quantization Codebook Design And Vq-Based Digital Watermarking Embedding Technology Study

As an effective data compression method, vector quantization (VQ) technology has been widely used in image compression and speech coding with it's especial merits of high compression rate and easy decoding arithmetic. Although VQ technology has been developed continually, the compression standard of VQ has not been build at present. The primary reason is that the standard of codebook has close pertinence with coding object, in different applications, the codebook structure, length and the dimensions of vector are different. So we must do more research work on the codebook design. With the increasing of VQ digital productions, how to protect the copyright of those productions efficiency has been an exact question. This paper has studied on the codebook design subject and the VQ productions' copyright protection. The works of this paper mainly include two parts.In the first part, for the original codebook selection problem in codebook design, this paper proposes a self-adaptive original codebook design algorithm to adapt the training vectors, then used this algorithm to improve on classical LBG codebook design algorithm and based on self-organizing feature map neural network codebook design method. The experimental result shows the improved algorithms can enhance the codebook quality and shorten the time of codebook design.The second part proposes two VQ-based digital watermarking embedding algorithms for the problem of VQ production's copyright production. In algorithm one, using authority key, watermarking data is embedded in codebook to prevent lawless users' tortuous actions. The algorithm two is based on algorithm one and improved it to track legal users' modificative actions on the VQ productions.