Audio Features Dedicated to the Detection of Four Basic Emotions

Grekow, Jacek

doi:10.1007/978-3-319-24369-6_49

Cited by 13 publications

(8 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Although there are some works that are focused only on prediction like [11], [29], [14] or [7], and that are some others centered only on solving classification problems like [28], [13], [44]; there are also several works that make predictions and later extend their systems to achieve classifications, such as [37] or [2]. Generally, these works that include prediction and classification use datasets annotated in dimensional models, in which a valence and arousal (V/A) coordinate system is established [34].…”

Section: Related Workmentioning

confidence: 99%

“…It can be observed that for some works no details on data balancing are incorporated, especially in prediction systems (see Table 1) in which it has no major relevance. Regarding classification systems (see Table 2), the data of the works [28], [13] and [44] are balanced and do not require any special treatment. Instead, [27] presents unbalance data, and although there is no detail of the strategy used to balance the data, apparently some songs are removed from the original dataset.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Emotional classification of music using neural networks with the MediaEval dataset

Medina

Beltrán

2020

Pers Ubiquit Comput

View full text Add to dashboard Cite

The proven ability of music to transmit emotions provokes the increasing interest in the development of new algorithms for music emotion recognition (MER). In this work, we present an automatic system of emotional classification of music by implementing a neural network. This work is based on a previous implementation of a dimensional emotional prediction system in which a Multilayer Perceptron (MLP) was trained with the freely available MediaEval database. Although these previous results are good in terms of the metrics of the prediction values, they are not good enough to obtain a classification by quadrant based on the valence and arousal values predicted by the neural network, mainly due to the imbalance between classes in the dataset. To achieve better classification values a pre-processing phase was implemented to stratify and balance the dataset. Three different classifiers have been compared: Linear Support Vector Machine (SVM), Random Forest and MLP. The best results are obtained with the MLP. An averaged F-measure of 50% is obtained in a four-quadrant classification schema. Two binary classification approaches are also presented: one vs. rest (OvR) approach in four-quadrants and binary classifier in valence and arousal. The OvR approach has an average F-measure of 69%, and the second one obtained F-measure of 73% and 69% in valence and arousal respectively. Finally, a dynamic classification analysis with different time windows was performed using the temporal annotation data of the MediaEval database. The results obtained show that the classification F-measures in four quadrants are practically constant, regardless of the duration of the time window. Also, this work reflects some limitations related to the characteristics of the dataset, including: size, class balance, quality of the annotations and the sound features available.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Emotional classification of music using neural networks with the MediaEval dataset

Medina

Beltrán

2020

Pers Ubiquit Comput

View full text Add to dashboard Cite

show abstract

“…Karar ağaçları yorumlanması kolay, veri yapılarını ifade etmede başarılı, hızlı, ve güvenli olması nedeniyle yaygın kullanılan makine öğrenme algoritmalarındandır. Deneysel çalışmalar çoğu zaman Weka da gerçekleştirildiği için karar ağacı olarak J.48 algoritması [29] kullanılmaktadır [21]. Bu algoritma çalışmalarda sık karşımıza çıkan C4.5 [30] algoritmasına dayanır ki bu algoritma da ağaç oluşturma algoritması ID3'ün geliştirilmiş versiyonudur.…”

Section: Karar Ağaçlarıunclassified

Emotion Recognition From Different Types of Music From Different Cultures

Hizlisoy¹,

Tüfekçi²

2020

Çukurova Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi

View full text Add to dashboard Cite

Bu çalışmada, klasik makine öğrenme yöntemleri farklı kültürlere ait farklı türdeki müziklerden oluşmuş veri tabanları üzerinde duygu tanıması yapmak için kullanılmışlardır. Bu veri tabanlarında bulunan müziklerden öznitelik çıkarmak için çalışmalarda yaygın olarak kullanılan araçlar tercih edilmiştir. Çıkarılan bütün özniteliklere korelasyon tabanlı öznitelik seçme yöntemi uygulanmıştır. Makine öğrenmesi yöntemleri olarak Bayes Ağları, Sıralı Minimal Optimizasyon, Lojistik Regresyon ve Karar Ağaçları kullanılmıştır. Öznitelik seçim işlemi sonrasında kalan özniteliklere Bayes Ağları yöntemi uygulandığında, Türkçe Duygusal Müzik Veri Tabanı için %94,35, Bi-Modal Veri Tabanı için %79,62 ve Soundtrack Veri Tabanı için ise %75,45 tanıma oranı elde edilmiş ve karşılaştırılan sınıflandırıcılardan daha iyi sonuç alınmıştır. Daha sonra, araçlardan çıkarılan öznitelikler bir araya getirilmiş ve yine seçim işlemi yapılmıştır. Bu işlemden sonra ise, sırasıyla bu veritabanları için %95,96, %80,24 ve %82,72 tanıma oranları elde edilmiştir.

show abstract

“…• Cluster 1: passionate, enthusiastic, confident, bustling, noisy. Some works in which this library is applied are: in [25] for the classification of music by musical genre for a database of 120 songs, in [26] is used for the detection of 4 basic emotions through a categorical affective model that recognize emotions: happy, angry, sad, relaxed; additionally also the precision of the emotion recognition is validated, varying the features selection that are extracted and used in the classification model; and finally in [6] the library is used to detect the valence and arousal values in musical recordings, showing the importance of combining low-level features with highlevel features to achieve better results in emotional classifications of music. Table 5 presents a comparison between the three libraries selected in this article: Spotify API, jAudio and AcousticBrainz.…”

Section: Jmir/jaudiomentioning

confidence: 99%

High-Level Libraries for Emotion Recognition in Music: A Review

Medina

Baldassarri

2018

Communications in Computer and Information Science

View full text Add to dashboard Cite

This article presents a review of high-level libraries that enable to recognize emotions in digital files of music. The main objective of the work is to study and compare different high-level content-analyzer libraries, showing their main functionalities, focused on the extraction of low and high level relevant features to classify musical pieces through an affective classification model. In addition, there has been a review of different works in which those libraries have been used to emotionally classify the musical pieces, through rhythmic and tonal features reconstruction, and the automatic annotation strategies applied, which generally incorporate machine learning techniques. For the comparative evaluation of the different high-level libraries, in addition to the common attributes in the chosen libraries, the most representative attributes in music emotion recognition field (MER) were selected. The comparative evaluation enables to identify the current development in MER regarding high-level libraries and to analyze the musical parameters that are related with emotions.

show abstract

Audio Features Dedicated to the Detection of Four Basic Emotions

Cited by 13 publications

References 12 publications

Emotional classification of music using neural networks with the MediaEval dataset

Emotional classification of music using neural networks with the MediaEval dataset

Emotion Recognition From Different Types of Music From Different Cultures

High-Level Libraries for Emotion Recognition in Music: A Review

Contact Info

Product

Resources

About