Video detection method and device, apparatus and medium

Abstract

The invention discloses a video detection method. The method comprises the steps of obtaining a to-be-detected target video; performing feature extraction on each key frame in the target video to generate a target video fingerprint, the target video fingerprint comprising the image features of each key frame in the target video; according to the target video fingerprint and a reference video fingerprint corresponding to a reference video in a reference video library, determining the distance from each key frame in the target video to the reference video, wherein the reference video fingerprint comprises the image characteristics of each key frame in the reference video; and determining the similarity between the target video and the reference video according to the distance between each key frame in the target video and the reference video. Due to the fact that the local space features and the local frame arrangement time sequence information of all the key frames are reserved, the spatial repetitive inspection is more accurate, and the long and short video mutual inspection and the similar video time position positioning are supported. The invention further discloses a corresponding device, an apparatus and a medium.

Description

technical field [0001] The present application relates to the technical field of video detection, in particular to a video detection method, device, equipment and computer storage medium. Background technique [0002] Nowadays, video platforms need to use video detection technology to realize effective management of video resources. For example, video detection technology is used to measure the similarity between two videos, so as to realize video management services such as video deduplication and piracy detection. [0003] At present, the commonly used video detection technology is to measure whether the two videos are similar by comparing the distance of the video fingerprints of the two videos; and the video fingerprint is specifically obtained by extracting the key frame features of the video key frames, and then through the dimensionality reduction algorithm. Dimensionality reduction is performed on the features, and finally all the key frame features of the video are ...

AI Technical Summary

Problems solved by technology

[0004] In the above commonly used video detection techniques, the local image features of the video are obtained first, and then the global video features are obtained based on the aggregation or averaging of the local image features. In the process of obtaining the global video features, the local image features will be lost. As a result, video fingerprints are poor in describing video image features, and ultimately lead to low effectiveness of video retrieval based on such video fingerprints, which affects the progress of video management services.

Images