Video auditing method and device and electronic equipment

Abstract

The invention discloses a video auditing method, a video auditing device and electronic equipment. The video auditing method comprises the following steps: obtaining a to-be-audited image frame; identifying a target object in the to-be-audited image frame; and performing preset processing on the to-be-audited image frame according to the target object, and outputting the processed to-be-audited image frame. In the video auditing method, target object identification is carried out on a to-be-audited video in advance, preset processing is carried out on the to-be-audited image frame according tothe target object, and the to-be-audited image frame is output to an audit; in this way, an auditor can be reminded that the target object exists in the to-be-audited image frame, the auditor is prevented from ignoring important content in the to-be-audited image frame due to some special conditions, the auditing speed can be increased, and the auditing workload can be reduced to a certain extent.

Description

technical field [0001] The present disclosure relates to the field of image processing, in particular to a video review method, device and electronic equipment. Background technique [0002] At present, there are a lot of short video content, and the content of the short video needs to be reviewed to supervise the video content. However, the current review is generally a combination of machine review and manual review. Generally, problematic videos classified after machine review will be transferred to manual review and tagged. Manual review generally extracts key frames from a video, and reviewers watch the key frames. , label the keyframes to determine whether the video violates the rules. However, usually each video needs to extract 8 key frames, and reviewers need to review 8 frames of pictures one by one and label them. Since reviewers cannot know the type of violation of the picture, they need to carefully observe the picture before labeling, which is time-consuming a...

AI Technical Summary

Problems solved by technology

However, usually each video needs to extract 8 key frames, and reviewers need to review 8 frames of pictures one by one and label them. Since reviewers cannot know the type of violation of the picture, they need to carefully observe the picture before labeling, which is time-consuming and laborious. In the state of reviewing for a long time, it is easy to miss the violating video

Images