Video watermark identification method and identification device thereof

Abstract

The invention provides a video watermark identification method and a video watermark identification device thereof. The video watermark identification method comprises the steps of: acquiring video frames of a video, and selecting one video frame as a comparison video frame; taking the video frame separated from the comparison video frame by a preset number of frames as a current video frame; identifying to obtain similar pixel points according to pixel values of pixel points at the same positions of the current video frame and the comparison video frame; removing non-similar pixel points except the similar pixel points from the current video frame, and taking the current video frame without the non-similar pixel points as a new comparison video frame; returning to the above steps until the last current video frame of the video is compared; acquiring adjacent pixel points adjacent to the similar pixel points, and acquiring a watermark range in the video according to the similar pixel points and the adjacent pixel points; and carrying out blocking processing on the watermark range to obtain at least one blocked watermark region. The video watermark identification method has the advantage that the watermark recognition speed is increased.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a video watermark recognition method and a recognition device thereof. Background technique [0002] With the rapid development of digital media technology and computer technology, digital media such as video can be spread rapidly through the Internet. In order to protect the copyright of the video, the watermark is usually embedded when the video is produced, and the video watermark is played simultaneously when the video is played. However, when the video is played, since the content of the video watermark and the original image of the video are compressed together, the watermark will be displayed while playing the original video. Therefore, a method for removing the watermark in the video is currently needed. [0003] In the prior art, the SIFT algorithm or neural network recognition algorithm is usually used to identify the watermark in the video, but the above-mentioned p...

AI Technical Summary

Problems solved by technology

[0004] First, the calculation speed is slow, and each calculation takes several minutes or even hours of calculation time, and the user cannot use other applications during this period, so it will bring a lot of inconvenience to the user;

[0005] Second, the complexity of the algorithm is high, so it needs to consume more computing resources, especially the deep learning technology needs to use dedicated GPU resources to accelerate the learning process;

[0006] Third, the scalability is insufficient. Technologies such as pattern recognition and deep learning need to prepare a large amount of training data in advance, and the training results and the selection of samples have a relatively strong correlation, so they can only be applied to limited types of scenarios.

Images