Research On APP Design Of Video Background Music Recommendation Based On Multimodal Sentiment Analysis

With the development of digital media technology,video is widely used to record our daily activities and important moments.Selecting a background music that matches the characteristics of video is very important for the expression of video content and sentiment rendering.However,selecting suitable background music for videos requires users to master certain video post-production skills and music professional related knowledge,and it is very difficult and time-consuming for users to find suitable background music from a large number of music collections.In order to realize the background music recommendation matching with the features of video,this paper launched the design research of the video background music recommendation algorithm and its application:Firstly,this paper implemented a model of multi-modal sentiment analysis based on deep learning.The model modeled the fine-grained structure of image modal information and voice modal information at the sub-word level,and generated a representation of fused multi-modal information.The feature vector representation of multi-modal video used for sentiment classification can well extract the sentiment features of the video.On the basis of extracting the sentiment features of the video,the features of color,light,motion,etc.of the video were extracted,while the rhythm,frequency,timbre and other features of the music were extracted.Multiple-type Latent Semantic Analysis(M-LSA)was used to model the content correlation between video features and music features,and the potential content relationship between music representation and video representation fused with sentiment features was captured.According to the matching degree,the ranking list of recommended background music was generated.Finally,the algorithm of video background music recommendation system was integrated into the product design of APP.And the target user needs analysis was completed through the research of the target user of the video background music recommendation APP;the core task framework and functional module design of the application was completed according to the user needs analysis;Based on the application of sentiment design theory in the field of interactive product design,the interface design and interactive design of the video background music recommendation APP were carried out.This paper provides an efficient auxiliary tool for video production.A product that can effectively complete the recommendation of video background music was realized.Among the massive music,it helps users to directionally select the background music that conforms to the video atmosphere and video content in the recommended range.