Audio Classification in Carnatic Music
Audio Classification in Carnatic Music
Embedded Files
I developed an audio classification system to categorize Carnatic vocal recordings into Vivadhi or Avivadhi Ragas using a K-Nearest Neighbors (KNN) approach with handcrafted features such as Spectral Centroid, RMS Energy, and Chroma Frequency. To enhance performance, I built a Convolutional Neural Network (CNN) for automated feature extraction, achieving validation accuracies of 66.66% with KNN and 75% with the CNN model optimized using Nadam. This research was accepted and presented at the 6th International Conference on MIND (NIT Goa) and will be published in Springer’s Communications in Computer and Information Science (CCIS) series.
I developed an audio classification system to categorize Carnatic vocal recordings into Vivadhi or Avivadhi Ragas using a K-Nearest Neighbors (KNN) approach with handcrafted features such as Spectral Centroid, RMS Energy, and Chroma Frequency. To enhance performance, I built a Convolutional Neural Network (CNN) for automated feature extraction, achieving validation accuracies of 66.66% with KNN and 75% with the CNN model optimized using Nadam. This research was accepted and presented at the 6th International Conference on MIND (NIT Goa) and will be published in Springer’s Communications in Computer and Information Science (CCIS) series.
Flowchart explaining the whole approach of the work
Results
Page updated
Google Sites
Report abuse