基于前向神经网络的音乐情感识别

Recognition of Music Emotion Based on Forward Neural Network

情感是音乐的固有属性。本文通过对基于音乐多特征的比较,最终选择旋律走势作为情感特征,介绍了前向神经网络运用于音乐情感识别的基本原理,对使用前向神经网络(Feed-forward neural network)进行音乐情感识别的结果进行分析。实验结果显示:前向神经网络在音乐情感识别领域有着较高的准确性,能够有效地对音乐情感进行识别。

The most common way to retrieve and describe music is emotional words.Emotion is an inherent attribute of music.Based on the comparison of multi-features based on music,the paper finally selects the melody trend as the emotional characteristics,introduces the basic principles of the forward neural network used in recognition of music emotion,and analyzes the result that uses feed-forward neural network to identify the emotion of music.The experimental results show that the forward neural network has a high accuracy in the field of emotional emotion recognition,and can effectively identify the emotion of music.