Bangla Music Genre Classification using Deep Learning and Machine Learning Techniques

Abstract

The rapid growth of digital music libraries and the volume of Bangla music has increased the need for efficient and automatic methods to categorize Bangla musical genres. This study provides a thorough examination of the development and implementation of a new Bangla Music Genre Classification model. The proposed model uses advanced machine learning techniques to analyze key elements of Bangla music compositions, allowing for accurate and efficient genre identification. The approach in this study extracts essential audio features including tempo, spectral centroid, chroma frequency, spectral rolloff, RMSE, spectral bandwidth and MFCC, from a complex dataset covering various Bangla music genres. A modern deep learning algorithm, specifically a convolutional neural network, is trained on carefully annotated data and uses these features as input. The research highlights the importance of tailoring feature extraction methods and model designs to the unique qualities of Bangla music. The model aims to improve the accuracy and cultural relevance of automated genre identification by addressing the challenges associated with the diversity within Bangla music. The model is thoroughly tested using benchmark datasets and its performance is compared with existing genre classification methods. The outcomes show how adaptable and effective the suggested method is, especially for Bangla music. This research provides a comprehensive overview of creating and applying a model for classifying Bangla music genres, offering a reliable and culturally aware solution for accurate genre classification in Bangla music. It not only addresses the technical challenges of automated genre identification but also recognizes the unique cultural aspects inherent in Bangla music.