Research Of Robust Video Watermarking Based On Quantization Modulation

The digital transformation of multimedia information and the popularization of network will undoubtedly provide a great convenience for people.Digital video is a popular multimedia form.However,due to its easy to copy and disseminate rapidly,video piracy and infringement cases have occurred,copyright protection is worrying.Digital video watermarking method provides a comprehensive and reliable solution for copyright protection,and has been developed into an important content of digital watermarking and information security.In this paper,we propose an improved watermark embedding algorithm based on angle tangent quantization for uncompressed digital video.The main contents are as follows:(1)Because of the large computational complexity and high computational complexity,only the key frames are selected and the key frames are obtained according to the key frame extraction method.The key frame extraction method of this paper is the result of the combination of two methods,namely,discrete cosine transform and Gaussian mixture model,discrete wavelet transform and human visual system,which are selected as the key frame from the video content and characteristic.(2)On the basis of the key frame,the significant point selection algorithm is adopted to get the significant point as the optimal watermark embedding position.The significant point selection algorithm in this paper makes use of the just noticeable difference model in the wavelet domain and makes the selected significant points have good masking properties and low correlation and low error rate by means of scrambling and blocking techniques.(3)After selecting key frames and significant points,this paper proposes a watermark embedding algorithm based on angle tangent quantization based on the previous contents.The algorithm improves angle discontinuity in angular quantization index modulation,and studies the angle tangent,and designs the quantization function,watermark embedding rule and the modification amount of wavelet coefficients.Experiments show that the proposed algorithm has obvious advantages over other algorithms in watermark imperceptibility and robustness.