Education
2022-Present
Ph.D. Computer Science
University of Windsor2018-2021
M.Sc. Computer Science _ Data Mining
Amirkabir University of Technology (Tehran Polytechnic)Thesis Title: "Persian modal music (Dastgah) Detection using Deep Neural Networks "
Abstract:
In this research, several deep neural networks are implemented to recognize Persian modal music in seven high correlated categories. The best model which achieved 92 percent overall accuracy is using an architecture inspired by Autoencoder including BiLSTM and BiGRU layers. This model is trained by using Nava dataset which has 1786 records and up to 50 hours of worth of music played solo by Kamanche, Tar, Setar, Ney, and Santoor(Dulcimer). Features that have been studied through this research, contain MFCC, Chroma feature and Mel spectrogram. The results indicate that MFCC carries more valuable information about Persian modal music than other features. Moreover, The architecture which is inspired by autoencoders is powerful in distinguishing high correlated data like Persian modal music. It also shows that because of the precise order in Iranian Dastgah Music, Bidirectional Recurrent networks are more efficient than any other network that has been implemented in this research. (Github link)