Fast video detection method and device based on video fingerprints

Abstract

The invention discloses a fast video detection method and device based on video fingerprints, which can identify the consistency between simply edited video and a homogenous video, thereby automatically controlling an infringing video which is repeatedly uploaded by a user, and reducing labor cost. The technical scheme provides the method which comprises the following steps of decoding an input video to obtain YUV (luma and chroma) components; performing dimension division for the Y components of the kth frame after the video is converted, obtaining an N*M array, wherein the dimension of each frame is N*M; calculating a center-of-mass gradient direction of the array for each block in the array, and obtaining a three-dimensional fingerprint vector of the kth frame; obtaining a fingerprint sequence according to the calculated three-dimensional fingerprint vector of the kth frame; matching the fingerprint sequence with a pre-extracted standard sequence vector of an object video, and identifying whether the input video and the target video are homogenous.

Description

technical field [0001] The invention relates to a method for detecting whether videos are of the same source, in particular to a method and a device for detecting whether videos are of the same source based on video files. Background technique [0002] With the rapid development of information networks, the number of users of online video services has also become very large. Providers of online video services have provided users with a variety of service models. Among them, user upload service (UGC), users can upload videos independently The video is uploaded to the platform of the video service provider for other users to watch, and the corresponding advertising commission can be obtained from the video service provider. Therefore, the user scale of the UGC business is very large, and the service provider also faces many difficult problems in the operation process , the most prominent problem is that the videos uploaded by users infringe the copyrights of others. Many users...

AI Technical Summary

Problems solved by technology

[0002] With the rapid development of information networks, the number of users of online video services has also become very large. Providers of online video services have provided users with a variety of service models. Among them, user upload service (UGC), users can upload videos independently The video is uploaded to the platform of the video service provider for other users to watch, and the corresponding advertising commission can be obtained from the video service provider. Therefore, the user scale of the UGC business is very large, and the service provider also faces many difficult problems in the operation process , the most prominent problem is that the videos uploaded by users infringe the copyrights of others. Many users upload works that infringe the copyrights of others. After the service provider deletes the video files, the users re-upload the video files to the platform after simple editing. Therefore, service providers need to invest a lot of manpower to monitor the legality of videos uploaded by users to avoid risks caused by infringement

Images