Novel favorite music classification using EEG-based optimal audio features selected via KDLPCCA

Sawata, Ryosuke; Ogawa, Takahiro; Haseyama, Miki

doi:10.1109/icassp.2016.7471777

Cited by 7 publications

(7 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since CCA needs the same number of samples from two heterogeneous sets, the lengths of an audio segment and an overlapping segment to calculate the mean and standard deviation of audio features were also in the same as those of EEG. Furthermore, we divided all of the audio feature vectors per musical piece in order to prevent [52] each subject overfitting caused by learning similar vectors extracted from the same musical piece, i.e., we employed a leave-trial-out validation.…”

Section: Results Of Classification Of Favorite Musical Piecesmentioning

confidence: 99%

“…CCA can extract a correlation via canonical variates obtained from a pair of multivariate datasets by maximizing a linear correlation. In [52], we previously proposed CCA-based audio feature selection, by which audio features suitable for individual music preference are selected. However, Yeh et al [53] reported that canonical variates projected by applying CCA to heterogeneous sets of features show better discriminative performance than that of original features if the heterogeneous sets have semantic relevancy.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

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

Sawata

Ogawa

Haseyama

2019

IEEE Trans. Affective Comput.

Self Cite

View full text Add to dashboard Cite

A novel audio feature projection using Kernel Discriminative Locality Preserving Canonical Correlation Analysis (KDLPCCA)-based correlation with electroencephalogram (EEG) features for favorite music classification is presented in this paper. The projected audio features reflect individual music preference adaptively since they are calculated by considering correlations with the user's EEG signals during listening to musical pieces that the user likes/dislikes via a novel CCA proposed in this paper. The novel CCA, called KDLPCCA, can consider not only a non-linear correlation but also local properties and discriminative information of each class sample, namely, music likes/dislikes. Specifically, local properties reflect intrinsic data structures of the original audio features, and discriminative information enhances the power of the final classification. Hence, the projected audio features have an optimal correlation with individual music preference reflected in the user's EEG signals, adaptively. If the KDLPCCA-based projection that can transform original audio features into novel audio features is calculated once, our method can extract projected audio features from a new musical piece without newly observing individual EEG signals. Our method therefore has a high level of practicability. Consequently, effective classification of user's favorite musical pieces via a Support Vector Machine (SVM) classifier using the new projected audio features becomes feasible. Experimental results show that our method for favorite music classification using projected audio features via the novel CCA outperforms methods using original audio features, EEG features and even audio features projected by other state-of-the-art CCAs.

show abstract

Section: Results Of Classification Of Favorite Musical Piecesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

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

Sawata

Ogawa

Haseyama

2019

IEEE Trans. Affective Comput.

Self Cite

View full text Add to dashboard Cite

show abstract

“…This resulted in two participants being excluded at this stage. The remaining twenty participants had a mean age of 22 (range [19][20][21][22][23][24][25][26][27][28][29][30]. Nine of these participants were female.…”

Section: Participantsmentioning

confidence: 99%

“…It has been demonstrated that it is possible to recognise emotions from the EEG via a variety of techniques [14], [15], [16], and that this may be done in real-time for use in music-therapy [17]. It has also been demonstrated that it is possible to recognise emotions from the EEG during music listening tasks [18] and from a combination of EEG and acoustic features [19], [20]. However, these approaches use a population based classification approach, in which the same classification model is applied to all the study participants.…”

Section: Introductionmentioning

confidence: 99%

Personalised, Multi-Modal, Affective State Detection for Hybrid Brain-Computer Music Interfacing

Daly

Williams

Malik

et al. 2020

IEEE Trans. Affective Comput.

View full text Add to dashboard Cite

Brain-computer music interfaces (BCMIs) may be used to modulate affective states, with applications in music therapy, composition, and entertainment. However, for such systems to work they need to be able to reliably detect their user's current affective state. We present a method for personalised affective state detection for use in BCMI. We compare it to a population-based detection method trained on 17 users and demonstrate that personalised affective state detection is significantly (p < 0.01) more accurate, with average improvements in accuracy of 10.2 % for valence and 9.3 % for arousal. We also compare a hybrid BCMI (a BCMI that combines physiological signals with neurological signals) to a conventional BCMI design (one based upon the use of only EEG features) and demonstrate that the hybrid design results in a significant (p < 0.01) 6.2 % improvement in performance for arousal classification and a significant (p < 0.01) 5.9 % improvement for valence classification.Index Terms-EEG, GSR, Affective state detection, BCMI, Personalised affective state detection. ! • I. Daly is with the Brain-computer interfaces and Neural Engineering laboratory, School

show abstract

“…Many methods use biological signals (e.g., brain waves and heart rate) to extract unique features for each user [14]- [19]. However, these approaches put a physical burden on the users since biological signals are usually obtained by a device attached to the body.…”

Section: Introductionmentioning

confidence: 99%

Accurate Estimation of Personalized Video Preference Using Multiple Users' Viewing Behavior

Ito

Ogawa

Haseyama

2018

IEICE Trans. Inf. & Syst.

Self Cite

View full text Add to dashboard Cite

Yoshiki ITO †a) , Student Member, Takahiro OGAWA †b) , and Miki HASEYAMA †c) , Members SUMMARY A method for accurate estimation of personalized video preference using multiple users' viewing behavior is presented in this paper. The proposed method uses three kinds of features: a video, user's viewing behavior and evaluation scores for the video given by a target user. First, the proposed method applies Supervised Multiview Spectral Embedding (SMSE) to obtain lower-dimensional video features suitable for the following correlation analysis. Next, supervised Multi-View Canonical Correlation Analysis (sMVCCA) is applied to integrate the three kinds of features. Then we can get optimal projections to obtain new visual features, "canonical video features" reflecting the target user's individual preference for a video based on sMVCCA. Furthermore, in our method, we use not only the target user's viewing behavior but also other users' viewing behavior for obtaining the optimal canonical video features of the target user. This unique approach is the biggest contribution of this paper. Finally, by integrating these canonical video features, Support Vector Ordinal Regression with Implicit Constraints (SVORIM) is trained in our method. Consequently, the target user's preference for a video can be estimated by using the trained SVORIM. Experimental results show the effectiveness of our method. key words: multiview approach, spectral embedding, canonical correlation analysis, video preference, viewing behavior *

show abstract

Novel favorite music classification using EEG-based optimal audio features selected via KDLPCCA

Cited by 7 publications

References 24 publications

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

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

Personalised, Multi-Modal, Affective State Detection for Hybrid Brain-Computer Music Interfacing

Accurate Estimation of Personalized Video Preference Using Multiple Users' Viewing Behavior

Contact Info

Product

Resources

About