3D video copyright comprehensive protection method based on deep reinforcement learning optimization

• Main Authors: ZHU ZIQI, WU YALIAN, HUANG MENGBIAO, WU CHENGCHENG, HOU JIAN
• Format: Patent
• Language: Chinese; English
• Published: 24.01.2020
• Subjects: CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses a 3D video copyright comprehensive protection method based on deep reinforcement learning optimization. The 3D video copyright comprehensive protection method specifically comprises the following steps: establishing a WMNET learning framework for 2D image copyright protection on the basis of a general reinforcement learning framework; training a WMNET framework; registeringthe copyright of the depth map; identifying the copyright of the depth map; and synthesizing the 3D video containing the copyright information. According to the method, an adaptive domain is designedfor a 2D image, a detailed algorithm or expert knowledge for each attack is not needed to confront the 2D image, and the invisibility of the watermark can be freely controlled. According to the method, for depth map protection, original image data is not modified any more, and the distortion degree of the synthesized 3D video after watermark embedding is effectively reduced. According to the method, the contradiction betw