Deep Learning-Based Classification of Indian Classical Music Based on Raga

Classical music is an integral part of Indian culture’s lineage. The raga serves as the foundation for all Indian classical music, including both Hindustani and Carnatic music. The term “raga” is typically used to refer to the melodic structure in Indian classical music. Traditional approaches to the classification of ragas are time consuming and inefficient. In this work, a convolutional neural network was proposed for raga classification. The proposed convolutional neural network was trained on 70 classes of ragas. The proposed convolutional neural network takes the spectrograms of the audio note and identifies the raga of the note. The proposed convolutional neural network model achieved a precision of 97.9% on raga classification. The performance of the proposed convolutional neural network was compared with the traditional classification techniques using standard performance metrics such as precision, recall, F1 score, and AUC-ROC. The comparison results show that the classification performance of the proposed convolutional neural network was superior to the state-of-the-art machine learning techniques.