基于AdaBoost的音乐情绪分类

王磊; 杜利民; 王劲林

doi:10.3724/SP.J.1146.2006.00120

基于AdaBoost的音乐情绪分类

doi: 10.3724/SP.J.1146.2006.00120

计量
- 文章访问数: 3880
- HTML全文浏览量: 72
- PDF下载量: 1722
- 被引次数: 0
出版历程
- 收稿日期: 2006-01-24
- 修回日期: 2006-05-17
- 刊出日期: 2007-09-19

Mood Classification of Music Using AdaBoost

摘要

摘要: 随着流媒体应用的蓬勃兴起，音频信号的自动分类开始成为工程与学术关注的热点之一。根据音乐信号对乐曲表现的情绪进行分类，由于涉及音乐信号的社会属性和自然属性的综合表征与模糊分类，因此处理方法相应需要在各种传统表征与分类方法的基础上进行机制筛选与架构优化。该文探讨了在AdaBoost算法，K-L变换和GMM模型的基础上构造弱分类器的方法，采用多层分类器结构，成功地实现了对音乐信号进行情绪分类。初步的实验对163首歌曲进行平静(Calm)，悲伤(Sad)，激动(Exciting)以及愉悦(Pleasant)4种类别的分类，训练集和测试集的分类准确率分别达到97.5%和93.9%，展示了这种方法的可行性和进一步发展的潜在价值。
- AdaBoost;音乐;情绪;音色;节奏;分类;K-L变换;多层分类器
Abstract: With fast development and boosting of stream media applications, automatic classification of audio signals becomes one of the hotspots on research and engineering. Since mood classification of music is involved with integrated representation and classification of social and natural properties of music, mechanism selection and architecture optimization should be implemented on the basis of different traditional music representations and classification methods. This paper discusses formation of weak classifiers in AdaBoost algorithm based on K-L transformation and GMM training and realizes mood classification of music with multi-layer classifier architecture. The experiments classify 163 songs into four mood classes: calm, sad, exciting and pleasant with 97.5% accuracy on training data and 93.9% accuracy on test data, which proves feasibility and potential value of this method.

HTML全文

参考文献(1)

Landy M. Emotions and music, how does music convey emotion? From learning to performing. Canadian Music Educator, 2004, 45(4): 28-33.[2]Schapire R E. A brief introduction to boosting, proceedings of the 16th International Joint Conference on Artificial Intelligence, Stockholm, 1999: 1401-1406.[3]Viola P and Jones M J. Robust real-time face detection[J].International Journal of Computer Vision.2004, 57(2):137-154[4]Guo G D, Zhang H J, and Li S Z. Boosting for content-based audio classification and retrieval: An evaluation. International Conference on Multimedia and Expo, Tokyo, 2001: 253-256.[5]Xiong Z Y and Huang T S. Boosting speech / non-speech classification using averaged mel-frequency cepstrum coefficients features. IEEE Pacific Rim Conference on Multimedia, Hsinchu, 2002: 573-580.[6]Ravindran S and Anderson D. Boosting as a dimensionality reduction tool for audio classification, Proceedings of IEEE International Symposium on Circuits and Systems, Vancouver, 2004: III-465-8.[7]Zhang T and Kuo J. Hierarchical system for content-based audio classification and retrieval. Proceedings of SPIEs Conference on Multimedia Storage and Archiving Systems III, Boston, 1998: 398-409.[8]Liu D, Lu L, and Zhang H J. Automatic mood detection from acoustic music data. Proceedings of International Symposium on Music Information Retrieval, Baltimore, 2003: 81-87.[9]Scheirer E D. Tempo and beat analysis of acoustic musical signals[J].Journal of Acoustic Society of America.1998, 103(1):588-601[10]Freund Y and Schapire R E. A decision-theoretic generalization of on-line learning and an application to boosting[J].Journal of Computer and System Sciences.1997, 55(1):119-139[11]边肇祺，张学工. 模式识别. 第二版.北京：清华大学出版社，2002: 第9章.[12]Huang X D, Acero A, and Hon H W. Spoken Language Processing: A Guide to Theory, Algorithm and System Development. 1st Edition. Upper Saddle River, NJ, Prentice Hall PTR, 2001, Chapter 5.

施引文献

资源附件(0)

访问统计