Musical Genre Classification Based on Deep Residual Auto-Encoder and Support Vector Machine
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Music brings pleasure and relaxation to people. Therefore, it is necessary to classify musical genres based onscenes. Identifying favorite musical genres from massive music data is a time-consuming and laborious task. Recent studies have suggested that machine learning algorithms are effective in distinguishing between variousmusical genres. However, meeting the actual requirements in terms of accuracy or timeliness is challenging. Inthis study, a hybrid machine learning model that combines a deep residual auto-encoder (DRAE) and supportvector machine (SVM) for musical genre recognition was proposed. Eight manually extracted features fromthe Mel-frequency cepstral coefficients (MFCC) were employed in the preprocessing stage as the hybrid musicdata source. During the training stage, DRAE was employed to extract feature maps, which were then used asinput for the SVM classifier. The experimental results indicated that this method achieved a 91.54% F1-scoreand 91.58% top-1 accuracy, outperforming existing approaches. This novel approach leverages deeparchitecture and conventional machine learning algorithms and provides a new horizon for musical genreclassification tasks.
Abstract
Music brings pleasure and relaxation to people. Therefore, it is necessary to classify musical genres based onscenes. Identifying favorite musical genres from massive music data is a time-consuming and laborious task. Recent studies have suggested that machine learning algorithms are effective in distinguishing between variousmusical genres. However, meeting the actual requirements in terms of accuracy or timeliness is challenging. Inthis study, a hybrid machine learning model that combines a deep residual auto-encoder (DRAE) and supportvector machine (SVM) for musical genre recognition was proposed. Eight manually extracted features fromthe Mel-frequency cepstral coefficients (MFCC) were employed in the preprocessing stage as the hybrid musicdata source. During the training stage, DRAE was employed to extract feature maps, which were then used asinput for the SVM classifier. The experimental results indicated that this method achieved a 91.54% F1-scoreand 91.58% top-1 accuracy, outperforming existing approaches. This novel approach leverages deeparchitecture and conventional machine learning algorithms and provides a new horizon for musical genreclassification tasks.
참고문헌 (21)
- R. Sarkar, "Music genre classification using frequency domain features", Proceedings of 2018 5th International Conference on Emerging Applications of Information Technology (EAIT), pp.1~4 (2018)
- C. Weiss, "Mid-level chord transition features for musical style analysis", Proceedings of 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp.341~345 (2019)
- J. H. Foleis, "Texture selection for automatic music genre classification", Applied Soft Computing, 89 (2020)[DOI]
- A. Vidwans, "Classifying cultural music using melodic features", Proceedings of 2020 International Conference on Signal Processing and Communications (SPCOM), pp.1~5 (2020)
- L. K. Puppala, "A novel music genre classification using convolutional neural network", Proceedings of 2021 6th International Conference on Communication and Electronics Systems (ICCES), pp.1246~1249 (2021)
- C. Chen, "Combined transfer and active learning for high accuracy music genre classification method", Proceedings of 2021 IEEE 2nd International Conference on Big Data, Artificial Intelligence and Internet of Things Engineering (ICBAIE), pp.53~56 (2021)
- J. L. Conceicao, "Applying supervised learning techniques to Brazilian music genre classification", Proceedings of 2020 XLVI Latin American Computing Conference (CLEI), pp.102~107 (2020)
- N. Pelchat, "Neural network music genre classification", Canadian Journal of Electrical and Computer Engineering, 43(3), pp.170~173 (2020)[DOI]
- J. S. Luz, "Ensemble of handcrafted and deep features for urban sound classification", Applied Acoustics, 175 (2021)[DOI]
- D. Taufik, "AutoVAT : an automated visual acuity test using spoken digit recognition with MEL frequency cepstral coefficients and convolutional neural network", Procedia Computer Science, 179, pp.458~467 (2021)[DOI]
- S. Wang, "Auto-encoder neural network based prediction of Texas poker opponent's behavior", Entertainment Computing, 40 (2022)[DOI]
- M. Seo, "A graph embedding technique for weighted graphs based on LSTM autoencoders", Journal of Information Processing Systems, 16(6), pp.1407~1423 (2020)[DOI]
- B. M. Aslahi-Shahri, "A hybrid method consisting of GA and SVM for intrusion detection system", Neural Computing and Applications, 27, pp.1669~1676 (2016)[DOI]
- A. C. Enache, "Anomaly intrusions detection based on support vector machines with an improved bat algorithm", Proceedings of 2015 20th International Conference on Control Systems and Computer Science, pp.317~321 (2015)
- Y. Chen, "Default prediction of automobile credit based on support vector machine", Journal of Information Processing Systems, 17(1), pp.75~88 (2021)[DOI]
- H. Dai, "Multiscale fuzzy entropy and PSO-SVM based fault diagnoses for airborne fuel pumps", Human-centric Computing and Information Sciences, 11 (2021)[DOI]
- "Deep residual learning for image recognition" (2015)[URL]
- "GTZAN Dataset - Music Genre Classification" (2022)[URL]
- Y. Singh, "Robustness of musical features on deep learning models for music genre classification", Expert Systems with Applications, 199 (2022)[DOI]
- Y. Yu, "Deep attention based music genre classification", Neurocomputing, 372, pp.84~91 (2020)[DOI]
- C. Liu, "Bottom-up broadcast neural network for music genre classification", Multimedia Tools and Applications, 80, pp.7313~7331 (2021)[DOI]