We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.SD

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Sound

Title: Comparing the Accuracy of Deep Neural Networks (DNN) and Convolutional Neural Network (CNN) in Music Genre Recognition (MGR): Experiments on Kurdish Music

Abstract: Musicologists use various labels to classify similar music styles under a shared title. But, non-specialists may categorize music differently. That could be through finding patterns in harmony, instruments, and form of the music. People usually identify a music genre solely by listening, but now computers and Artificial Intelligence (AI) can automate this process. The work on applying AI in the classification of types of music has been growing recently, but there is no evidence of such research on the Kurdish music genres. In this research, we developed a dataset that contains 880 samples from eight different Kurdish music genres. We evaluated two machine learning approaches, a Deep Neural Network (DNN) and a Convolutional Neural Network (CNN), to recognize the genres. The results showed that the CNN model outperformed the DNN by achieving 92% versus 90% accuracy.
Comments: 8 pages, 5 figures, 3 tables
Subjects: Sound (cs.SD); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
Cite as: arXiv:2111.11063 [cs.SD]
  (or arXiv:2111.11063v1 [cs.SD] for this version)

Submission history

From: Hossein Hassani [view email]
[v1] Mon, 22 Nov 2021 09:21:48 GMT (297kb,D)

Link back to: arXiv, form interface, contact.