Research On3D Video Watermarking Algorithm

With the daily popularization and rapid development of the computer networkcommunication and digital multimedia technology， the applications of data informationtransmission in a wide variety of digital product is more and more. At this stage,three-dimensional (3D) products have come into our lives. Among them,3D video services canprovide more realistic and immersive experiences to end users when compared to traditionaltwo-dimensional (2D) multimedia services, which is considered to be the next generation ofvideo system. In application and dissemination process, the video would probably be damagedbecause various unintentional or malicious attacks, which makes the copyright protection of3Dproducts imminent. How to use the unique feature information of3D video image to reasonableand effective research the3D video watermarking is a new task that is worth to explore andstudy deeply. This paper combined with the latest3D video Big Hero video deeply researchesthe3D video zero watermarking algorithm.The main work of this paper is as follows:Firstly, the basic details of3D video technology and digital watermarking technology werestudied, including eye stereo vision principle,3D video technology introduction,3D stereo videodisplay and the classification and evaluation of digital watermarking, the type of attack, and thebasic framework. On this basis, further introduction of zero watermark technology principle,registration and detection process of zero watermarking is underway, which is the preparationsfor the next step of putting forward3D video zero watermarking algorithm based on DCTtransform.Secondly, in this paper, according to the principle of zero watermarking algorithms weproposed a new algorithm which theoretical basis is the use of the DCT transform coefficients, called3D video zero watermarking algorithm based on DCT transform. The preprocessingalgorithms of watermarking is introduced in the algorithm, which including the Arnoldscrambling transformation, Logistic chaotic encryption algorithm was proposed. Utilizing thedifference of3D video views feature extraction value, we can structure and extract the3D videozero watermark. This algorithm can solve the conflict between robustness and invisibilityeffectively during embedding, and it can also improve the robustness and security.Thirdly, in this paper, we have made the simulation experiment of a3D video zerowatermarking algorithm based on DCT transform in the MATLAB7.0version with the new3Dvideo Big Hero as carrier, then carried out the attack experiment. After analysis of theexperimental results, the algorithm can resist some simple attack and make image distortion freeand be robust, simple and ease to use which shows that the robustness and invisibility of thealgorithm is good.Lastly, because in the extraction method of video feature value is inadequate in3D videozero watermarking algorithm based on DCT transform, and the anti attack capability is notparticularly significant. Due to the introduction of SVD singular value decomposition method inoriginal algorithm we realize the improvement of original algorithm. SVD singular valuedecomposition has the unique characters of stability, transpose invariance, shift invariance androtation invariance. In the digital watermarking algorithm，we can make full use of the singularvalue decomposition characteristics to improve the3D video zero watermarking algorithm basedon DCT transform, which can enhance its robustness. Through calculate the simulation results ofthe relevant data, the paper shows that the improved3D video zero watermarking algorithmsbased on DCT transform and SVD decomposition has a better advantage both in robustness,resistance to aggressive or stability.