Video data processing method and device, computer and readable storage medium

Abstract

The embodiment of the invention discloses a video data processing method and device, a computer and a readable storage medium, computer vision, deep learning and other technologies in the field of artificial intelligence can be used, and the method comprises the steps: obtaining a target video, and obtaining a to-be-detected image from the target video, identifying a prediction icon area in the to-be-detected image and a prediction icon type corresponding to the prediction icon area; intercepting a prediction icon area from the to-be-detected image to serve as a candidate area, and obtaining astandard icon corresponding to the prediction icon type; matching the candidate area with a standard icon, and if a matching result of the candidate area and the standard icon is a matching success result, determining that the standard icon exists in the target video; and dividing the target video into video categories corresponding to the standard icons. Accuracy of identifying icons in videos is improved.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a video data processing method, device, computer and readable storage medium. Background technique [0002] An icon is a graphic identification of TV media and network media, etc., and is generally located in the area around the four vertices of the video. The icon here specifically refers to a graphic rather than a text-based identification, for example, a media logo (logo) or a personal logo. When it is necessary to determine the publisher of the video, or to classify the video based on the publisher, it is necessary to identify the icon in the video. At present, the icons in the video are generally recognized manually, or the icons in the video are detected and recognized by using a convolutional neural network. In general, the number of videos to be recognized is large, and the duration of each video is variable. That is, there may be a long video, and icons may ...

AI Technical Summary

Problems solved by technology

At present, the icons in the video are generally recognized manually, or the icons in the video are detected and recognized by using a convolutional neural network. In general, the number of videos to be recognized is large, and the duration of each video is variable. That is, there may be a long video, and there may not be icons in every frame of the video, so that manual identification will consume a lot of time and energy

When the convolutional neural network is used to identify the position of the icon in the video, due to the strong generalization of the convolutional neural network, image blocks with similar semantics may be recognized as icons, resulting in poor icon recognition accuracy.

Images