Hierarchical Dirichlet Process Mixture Model for Music Emotion Recognition

Abstract

This study proposes a novel multi-label music emotion recognition (MER) system. An emotion cannot be defined clearly in the real world because the classes of emotions are usually considered overlapping. Accordingly, this study proposes an MER system that is based on hierarchical Dirichlet process mixture model (HPDMM), whose components can be shared between models of each emotion. Moreover, the HDPMM is improved by adding a discriminant factor to the proposed system based on the concept of linear discriminant analysis. The proposed system represents an emotion using weighting coefficients that are related to a global set of components. Moreover, three methods are proposed to compute the weighting coefficients of testing data, and the weighting coefficients are used to determine whether or not the testing data contain certain emotional content. In the tasks of music emotion annotation and retrieval, experimental results show that the proposed MER system outperforms state-of-the-art systems in terms of F-score and mean average precision. Index Terms-Discriminant method, hierarchical Dirichlet process mixture model, music annotation and retrieval, music emotion recognition.

A Survey on Autonomous Techniques for Music Classification based on Human Emotions Recognition

“…Self collected [17], [18], [19], [20], [21], [22] RECOLA [23] CAL500 [24], [25], [26] Magna Tag A Tune [27], [28] MSD [27], [29] AMG 1608 [30], [31], [32] MER60 [30], [33] DEAP [34], [35], [36] Mediaeval [37], [38], [39] GTZAN [16] Marsyas [40]…”

“…Low mean squared error [66] F-measure [18], [24], [25], [78], [107] Accuracy [19], [37], [38], [39], [48], [61], [67], [78], [99] , [101], [108] Specificity [61] AuC [27] Pearson Correlation Coefficient [68], [109] R 2 statistics [3], [15], [38], [40], [65], [79], Precision [24], [25], [26] , [78], [107] Recall [24], [26], [78], [107] RMSE [39] Equal error rate [77]…”

Guest Editorial: Challenges and Perspectives for Affective Analysis in Multimedia

“…Extensive numerical results showed that the dimensionality-reduced features lead to better classification accuracy, comparing to the original features. Wang et al [7] explored a non-parametric Bayesian approach for emotional characterization of music. To construct a discriminative latent space while capturing correlations between emotions, they used a hierarchical Dirichlet process (HDP) mixture and modified it by introducing linear discriminant idea into the sampling distributions of HDP latent topics.…”

“…The mission of the special section is to provide a view at the state-of-the-art in this domain and point to the potentially interesting future research directions. It therefore features articles concerning multimedia affective understanding from different perspectives, including predicting likability and emotions from users' responses [3], [4], content analysis for music affective understanding [5], [6], [7], visual content analysis for aesthetics and emotional understanding [8], [9], [10], and an application for sentiment analysis [11].…”

Novel Audio Feature Projection Using KDLPCCA-Based Correlation with EEG Features for Favorite Music Classification

“…R ecently, a large amount of music has become available to users, and efficient music information retrieval technology is therefore necessary to retrieve musical pieces desired by the users. Many methods related to genre classification [1], [2], [3], [4], [5], [6], [7], artist identification [8], [9], music mood classification [10], [11], [12], [13], instrument recognition [14], [15], [16], [17] and music annotation [18], [19] have been proposed to help users retrieve desired musical pieces. Although these methods can help users to retrieve desired musical pieces from music databases, retrieving desired musical pieces from enormous music databases is still a laborious task.…”