A Machine Learning Approach to Automatic Music Genre Classification

Silla, Carlos N.; Koerich, Alessandro L.; Kaestner, Celso A. A.

doi:10.1007/bf03192561

Cited by 41 publications

(14 citation statements)

References 23 publications

(12 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper we use the time-decomposition approach [22,24] with an "ALL"-features hybrid vector that encompasses 298 features, where: 40 from Inset-Onset Interval Histogram Coefficient descriptors [9], 30 from MARSYAS [26], 60 from Rhythm Histograms [14] and 168 from Statistical Spectrum Descriptors [14]). Therefore, for each music piece all these features are extracted from 30-second segments chosen from the beginning (B), middle (M) and end (E) parts of the music signal.…”

Section: System Overviewmentioning

confidence: 99%

“…Music genre classification is carried out considering each classifier isolated as well as the combination of their outputs through combination rules such as majority voting of the class labels (MAJ), summation (SUM) or multiplication (PROD) of the posterior probabilities of each class [12]. Given the "ALL"-features hybrid vector, a feature selection procedure based on Genetic-Algorithm (GA) is applied [24]. We refer to the resulting feature vector as the GA-selected feature vector.…”

Section: System Overviewmentioning

confidence: 99%

“…This is due to the fact that the music genre in one of the most frequent items used in search queries, as pointed by several authors [5,13,21]. Over the past decade this task has received growing attention every year which lead to the development of different types of feature descriptors [9,14,26] and classifications approaches [1,15,16,18,24]. Current research on automatic music genre classification suggests that the use of content-based approaches have reached a glass ceiling [15].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Improving automatic music genre classification with hybrid content-based feature vectors

Silla

Koerich

Kaestner

2010

Proceedings of the 2010 ACM Symposium on Applied Computing

Self Cite

View full text Add to dashboard Cite

Current research on the task of automatic music genre classification has been focusing on new classification approaches based on combining information from other sources than the music signal. The reason for this is that the use of contentbased approaches, i.e. using features extracted directly from the audio signal, seems to have reached a glass ceiling. In this work we show that by using different types of contentbased features together it is possible to substantially improve the classification accuracy. This is an interesting result as different types of content-based features aim, at a conceptual level, to capture the same type of information. In order to identify which types of content-based features are responsible for the predictive accuracy gain, we also used a feature selection (FS) approach based on a genetic algorithm (GA). The analysis of the results in two databases shows that the use of the GA for FS succeeds in selecting a representative subset without significant loss in accuracy. It also shows that all the different types of content-based features employed are important for the improvement of the accuracy in classifying music genres.

show abstract

Section: System Overviewmentioning

confidence: 99%

Section: System Overviewmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Improving automatic music genre classification with hybrid content-based feature vectors

Silla

Koerich

Kaestner

2010

Proceedings of the 2010 ACM Symposium on Applied Computing

Self Cite

View full text Add to dashboard Cite

show abstract

“…[17,[21][22][23]27]. A more detailed description and comparison of these works can be found in [39]. On the other hand, few works have dealt with feature selection.…”

Section: Problem Definition and Related Workmentioning

confidence: 99%

“…In our previous work we have employed the MARSYAS framework for feature extraction [39,40]. Such a framework extracts acoustic features from audio frames and aggregates them into high-level music segments [42].…”

Section: Feature Setsmentioning

confidence: 99%

A Feature Selection Approach for Automatic Music Genre Classification

Silla

Koerich

Kaestner

2009

Int. J. Semantic Computing

Self Cite

View full text Add to dashboard Cite

In this paper we present an analysis of the suitability of four different feature sets which are currently employed to represent music signals in the context of the automatic music genre classification. To such an aim, feature selection is carried out through genetic algorithms, and it is applied to multiple feature vectors generated from different segments of the music signal. The feature sets used in this paper, which encompass time-domain and frequency-domain characteristics of the music signal, comprise: short-time Fourier transform, Mel frequency cepstral coefficient, beat-related features, pitch-related features, inter-onset interval histogram coefficients, rhythm histograms and statistical spectrum descriptors. The classification is based on the use of multiple feature vectors and an ensemble approach, according to time and space decomposition strategies. Feature vectors are extracted from music segments from the beginning, middle and end parts of the music signal (time-decomposition). Despite music genre classification being a multi-class problem, we accomplish the task using a combination of binary classifiers, whose results are merged to produce the final music genre label (space decomposition). Experiments were carried out on two databases: the Latin Music Database, which contains 3,227 music pieces categorized into ten musical genres; the ISMIR'2004 genre contest database which contains 1,458 music pieces categorized into six popular western musical genres. The experimental results have shown that the feature sets have different importance according to the part of the music signal from where the feature vectors are extracted. Furthermore, the ensemble approach provides better results than the individual segments in most cases. For high-dimensional feature sets, the feature selection provides a compact but discriminative feature subset which has an interesting trade-off between classification accuracy and computational effort.

show abstract