Research On Content Based Video Copyright Protection

The rapid development of multimedia technology has greatly promoted the production of digital video,and the degree of freedom in the creation,dissemination,and storage of digital video has increased.Digital video brings convenience to people.At the same time,the legality of the copyright of multimedia video is becoming more and more prominent,and the problems faced in the maintenance of video copyright are becoming more and more complicated.The criminals use various advanced video editing and processing software to easily distort the original video and seriously damage the legitimate interests of video copyright owners.In the face of complex online video,effectively safeguarding the legal copyright of digital video is not only an important basis for creating a good Internet environment,but also a protection for every legitimate rights and interests of the Internet.It is also to further promote the fairness,justice and harmony of the Internet environment.The foundation of beautiful and orderly development.Generally in digital video copyright protection,a systematic digital copyright protection model not only involves verification of content,but also has high requirements for video preprocessing.Therefore,the realization of key technologies based on video copyright protection is the top priority in maintaining the legitimacy of digital video copyright.This article focuses on video shot segmentation,key frame extraction and video digital watermark embedding extraction in the key technologies of video copyright protection.The main contents include:(1)A shot segmentation method based on auto-regressive model and mutual information is proposed for shot segmentation.The HSV histogram feature vector is used to calculate the mutual information of two frames and convert it into similarity values.Then the auto-regression is established by calculating the similarity values.The model retrieves the decision value of the discriminant to obtain the adaptive threshold,and finally determines the boundary of the lens in combination with the generated threshold and the set frame time window to segment the video shot.The experiment verifies that the algorithm has a good detection effect and improves the overall performance of the shot segmentation.(2)In order to reduce the computational complexity of key frame extraction,a method based on mutual information is used to extract the key frames.A key frame extraction method based on the variance of mutual information is proposed,which firstly calculates the mutual information amount.The mean squared error is then compared with the given threshold.If the threshold is greater than the threshold,the key frame is selected according to the ratio of the number of frames in the shot.If it is less than the threshold,the first and last points in the shot are taken.The most critical frame of the frame,so as to achieve the extraction of key frames.The algorithm is relatively simple to implement and has a good effect of key frame extraction.(3)In order to better deal with the attacker's destruction of the embedded watermark information,an adaptive digital watermarking algorithm based o n HVS in the key frame of DCT is proposed.The watermark information is scrambled and the computational complexity and clutter are reduced as much as possible.Therefore,the watermark information is embedded and extracted.Experiments show that the algorithm has low computational complexity,accurate watermark detection,and good watermark protection.