Video style identification method and device and electronic equipment

Abstract

The invention discloses a video style identification method and device and electronic equipment, and the method belongs to the field of artificial intelligence, and the method comprises the steps: carrying out the analysis of a target video through a style analysis model, obtaining the style information of the target video, wherein the style information represents the high and low degree of the style of the target video, in this way, video recommendation can be more accurately carried out on the user through the identified degree of the video style. A first training sample for training the style analysis model is a style-related video acquired from a negative feedback video, and the negative feedback video represents a video in which the user is relatively low in interesting degree, so that the training sample acquired from the negative feedback video can better represent the actual demand of the user. In this way, the result meeting the user requirement can be analyzed through the style analysis model obtained through training of the training sample.

Description

technical field [0001] The invention relates to the field of video processing, in particular to a video style recognition method, device and electronic equipment. Background technique [0002] With the continuous development of the short video field, the content of short videos is also becoming more and more diverse. Some users have made some low-key videos in order to pursue clicks. It's a video of Sansu. For this kind of video, the public is very controversial. Some people don't like to watch this kind of video, and even feel very disgusted. However, some people like to watch it. Therefore, in order to provide targeted recommendations to users, it is necessary to identify low-key videos. [0003] The classification methods for video style in the prior art generally manually define low-key standards, and select training samples through manually defined standards. However, artificially defined low-key standards have strong limitations and cannot reflect real user needs. ...

AI Technical Summary

Problems solved by technology

[0003] The classification methods for video style in the prior art generally manually define low-key standards, and select training samples through manually defined standards. However, the artificially defined low-key standards have strong limitations and cannot reflect real user needs

[0004] In addition, in the prior art, a binary classification model is usually used to identify low-key videos, and the recognition results only include two situations, which cannot reflect the level of video style. Applying the recognition results of the binary classification model to information recommendation less effective

Images