A video classification method, device and computer-readable storage medium

Abstract

The present invention provides a video classification method, device, and computer-readable storage medium, comprising: extracting video pose data in a target video; matching the video pose data with seed pose data included in a preset pose retrieval library, and determining the video pose data The posture classification category corresponding to the posture data, the posture retrieval database includes the corresponding relationship between the posture data and the posture classification category; the posture classification category corresponding to the video posture data is determined as the classification category of the target video, and the present invention can pass the video posture data and the seed posture The data is matched in the attitude dimension to achieve the purpose of classifying the target video in the attitude dimension. The interference caused by information in other dimensions is reduced, thereby reducing the amount of data calculation and solving the problem of large calculation due to the calculation of feature similarity. In addition, through the matching method of pose data, the operation of frame-by-frame feature extraction on the video is avoided, and the efficiency of classifying the target video is improved.

Description

technical field [0001] The present invention belongs to the field of computer technology, and in particular, relates to a video classification method, a device and a computer-readable storage medium. Background technique [0002] In long videos, the characters shown in some plots often have a lot of exaggerated body movements in posture, and the video sections containing these contents are often relatively exciting video clips. , can efficiently and accurately find the video clips corresponding to these exaggerated gestures, and classify these video clips, which has important value and significance to the video business. [0003] In the prior art, many wonderful pictures will be displayed in the video, and the segments corresponding to these wonderful pictures are extracted and classified. Currently, a classification model based on the analysis of the features of static pictures is often used for processing. Specifically, the classification model consists of: Manually defin...

AI Technical Summary

Problems solved by technology

[0005] In view of this, the present invention provides a video classification method, device, and computer-readable storage medium, which solves the problems of large amount of data calculation and low processing efficiency in the current solution for video classification to a certain extent

Images