Method for counting video playing times

Abstract

The invention relates to the technical field of image processing technology and pattern recognition, and discloses a method for counting video playing times. The method comprises the following steps of: S1, defining a video image sequence F={f1,f2,f3,ellipsis,fN}, and extracting the feature vectors of key frames to obtain a feature vector set adv={V1,V2,ellipsis,VM} and a feature database adv_db={adv1,adv2,ellipsis,advL}; and S2, detecting a certain quantity of videos, and counting the video playing times according to the feature database. In the method, a video is presented by adopting a group of key features, and a function of automatically extracting key features from the video is realized by using a method for combining a multi-scale grid with a haar feature, so that image content matching of real-time videos and template videos is realized; and moreover, the playing times of video advertisements are counted by using a key feature frequency histogram accumulated integration method, and key feature vectors are short and are rapid for matching, so that the method is very suitable for searching and comparing large-scale video libraries, and is suitable for counting video playing times.

Description

Technical field [0001] The invention relates to the field of image processing technology and pattern recognition technology, in particular to a method for counting the number of times of video playback. Background technique [0002] Traditional video advertisement broadcast statistics are mainly realized manually, that is, personnel replay the video after the fact, manually record the duration, name, content, and type of the advertisement, and then summarize the information. This manual monitoring method is time-consuming and labor-intensive, and can no longer meet the needs of large-scale multimedia information statistical query. [0003] The intelligent advertising statistics technology is to realize the function of automatically counting the number of advertisements played through video analysis technology. This technology uses image processing and pattern recognition technology to monitor video advertisements. By monitoring real-time TV programs or program recording data, it r...

AI Technical Summary

Problems solved by technology

Because this technology uses multi-level feature retrieval technology, the template matching speed is improved to a certain extent, but because of the subsequent use of pixel-by-pixel matching method, the feature dimension is still relatively large, which is not conducive to large-scale matching, and pixel-by-pixel matching is not easy for pixel changes. It is more sensitive. If there are significant local changes in the image, such as the addition of floating subtitles and hanging corner advertisement information, the accuracy of advertisement matching will drop significantly

In addition, the advertising template video lacks automatic segmentation and extraction technology, which makes the template data relatively large, which is not conducive to large-scale advertising video data storage and search

Images