Deep Learning for Music Genre Classification: A case study of Thai music
Pasin Sawaengsawangarom, Suparoek Phongoen, Papis Wongchaisuwat · 2025
Classifying music genres plays an important role in music recommendation systems and information retrieval. Advanced deep learning model has provided promising results compared to other traditional methods. In this study, we explored two common deep learning models, namely convolutional neural networks, and recurrent neural networks, in classifying music genres. Mel-Frequency Cepstral Coefficients were extracted from the original audio data and used as features for training the models. We further verified the generalizability of models trained on publicly available dataset containing Western music to a custom-collected Thai music data. Our results highlighted both the potential and the current limitations of using deep learning for classifying diverse music cultures. This will contribute to the future development of more robust music genre classification systems.