Scene classification method, system and equipment combined with salient region detection, and medium

Abstract

The invention discloses a scene classification method, system and equipment combined with salient region detection, and a medium, and the method comprises the following steps: a short video image processing step: carrying out the frame extraction of a short video image, obtaining a plurality of video frames, and obtaining an image data set with a plurality of pieces of image information corresponding to the plurality of video frames according to the plurality of video frames; a non-salient region acquisition step: identifying a salient region and a non-salient region of each piece of picture information through salient region detection, and performing enhancement processing on the non-salient region; a classification and identification step: performing classification and identification on the non-salient region of each piece of picture information through a resnet-32 network structure to obtain an image scene probability of each piece of picture information; and an image probability processing step: calculating the scene probability of each scene category of the short video image according to the image scene probability of each piece of picture information. According to the invention, the identification accuracy and classification performance of the short video image can be improved.

Description

Technical field [0001] The present invention relates to the field of video processing, and Background technique [0002] With the rapid development of science and technology, communication technology has developed rapid development, and current communication technology has developed to the fifth generation communication --- 5G. The rapid development of 5G technology has brought new ways to internet information transmission - short video. Short video has become the most abundant, fastest and most popular way to spread information under social life. Under this rich month, the amount of data volume of various short video platforms rises sharply, how can we better use rich short video, and extract useful information, and data analysis has also become a new technical topic. . [0003] Unlike normal scene classification, short video scenarios are mainly facing two issues to be resolved: [0004] One is different from the image, only one scene will only be described in a pair image. lik...

AI Technical Summary

Problems solved by technology

[0005] Second, different from ordinary scene recognition, in the short video scene, the content of the surrounding environment information is relatively scarce, most of which are people occupying the main area, so how to accurately recognize the scene with a small amount of information is also a very big challenge

But the disadvantage of this method is: the video-based method can only give the subject scene of this video, and generally ignores other scenes that appear in the short video with a relatively short duration, resulting in a result that does not meet the requirements

But the disadvantage of this method is: but under the image-level conditions, the subject in the image is blocked by people, so the recognition accuracy may not be high

Images