Music recommendation method, device and system

Abstract

The embodiment of the invention provides a music recommendation method and device, an electronic device and a storage medium. The method specifically comprises the following steps: receiving video data and to-be-matched multi-segment music data, and inputting the video data and the music data into a pre-trained multi-modal matching network model for processing to obtain a video embedding vector ofthe video data and an audio embedding vector of each segment of music data; inputting the video embedding vector and the audio embedding vector into a pre-trained personalized sorting model for calculation to obtain a matching degree between the video data and each section of music data; and finally, outputting the music data of which the matching degree conforms to a preset standard as target music to be matched, so that the user can utilize the target music to carry out music matching on the video. The target music is obtained through objective calculation, and the result is objective and reliable, so that the matching degree between the dubbing music and the corresponding video work is effectively improved, and finally the satisfactory audio and video work is obtained.

Description

technical field [0001] The present disclosure relates to the technical field of data processing, and in particular to a music recommendation method, device, electronic equipment and storage medium. Background technique [0002] Short video content sharing and social apps have become the most popular application categories on the mobile Internet. Users can shoot, upload and share their short video works anytime and anywhere. Before uploading short video works, users can match the corresponding music background to the captured content, which can improve the rhythm and appeal of the video works. [0003] However, for ordinary users, choosing a suitable soundtrack for a short video is a relatively difficult problem. This is because short videos are mostly shot impromptu by users, with a variety of content, and ordinary users do not have professional music literacy, and have limited understanding of the rhythm and style of music, which leads to the matching of the soundtrack cho...

AI Technical Summary

Problems solved by technology

[0003] However, for ordinary users, choosing a suitable soundtrack for a short video is a difficult problem

This is because short videos are mostly shot impromptu by users, with a variety of content, and ordinary users do not have professional music literacy, and have limited understanding of the rhythm and style of music, which leads to the matching of the soundtrack chosen by ordinary users with the corresponding video works low degree

Images