Machine Learning Based Indo-Western Music Identification, Classification and Disease Healing
D C Shubhangi, Baswaraj Gadgay, Manjula Moolbharathi · 2023
Progress in availability of musical data over the last several years has given renewed importance to the task of categorizing musical genres. For easier access, we need an accurate index of them. Automatic music genre categorization is vital whenever dealing with large music library. Most recent efforts to categorize musical styles have used machine learning approaches. In this study, we utilized two data sets that included a wide range of genres. Deep Learning is used to train and categorize the system. A convolutional neural network is used for both classification and training. The Mel Frequency Cepstral Coefficient (MFCC) is now the most used tool for extracting audio features. Using the extracted feature vector, the proposed technique classifies songs into indo-western genres. In this paper, we offer a system for classifying and identifying Indo-Western musical styles. We thought about the GTZAN data collection for western music. One hundred samples, each lasting 30 seconds, have been compiled to represent each genre. We only included the top five genres from GTZAN dataset since there are five distinct types of Indian music. The research uses a neural network to identify indo-western hybrids and Indian classical compositions by feeding them tracks from each genre. In addition, if the classical raga is identified, the categorization of classical ragas associated with illness treatment is provided. It tells us which raga is effective against particular illness. We're demonstrating the physiological effects of each raga. In this paper, we implement machine learning algorithm for raga identification and assess its performance, contrasting the accuracy of a system that only uses MFCC as a feature with that of one that also makes use of pitch and Chroma information. Raga identification for therapeutic use is suggested using an algorithm based on machine learning. This technique extracts features using MFCC features, together with Pitch and Chroma data. Ragas was sorted using a K-Nearest Neighbors approach. The study is expanded to use an open-set method to determine the identities of many ragas (including Asvari, Bageshree, Bhairavi, Darbari, and Yaman). Diseases like heart disease, depression, autism, and addiction may all be treated by playing the appropriate raga, therefore it's important to know which one to utilize. Detailed explanation of the positive outcomes associated with raga.