KLASIFIKASI GENRE MUSIK MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK PADA DATASET MEL-SPECTROGRAM

WIMBASSA, MUHAMAD DWIRIZQY and Yusliani, Novi and Rizqie, M. Qurhanul (2024) KLASIFIKASI GENRE MUSIK MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK PADA DATASET MEL-SPECTROGRAM. Undergraduate thesis, Sriwijaya University.

	Text RAMA_55201_09021382126126.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (2MB) | Request a copy
	Text RAMA_55201_09021382126126_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (5MB) | Request a copy
	Text RAMA_55201_09021382126126_0008118205_0203128701_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (667kB)
	Text RAMA_55201_09021382126126_0008118205_0203128701_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (467kB) | Request a copy
	Text RAMA_55201_09021382126126_0008118205_0203128701_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_55201_09021382126126_0008118205_0203128701_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (839kB) | Request a copy
	Text RAMA_55201_09021382126126_0008118205_0203128701_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (524kB) | Request a copy
	Text RAMA_55201_09021382126126_0008118205_0203128701_06.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (206kB) | Request a copy
	Text RAMA_55201_09021382126126_0008118205_0203128701_07_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (208kB) | Request a copy
	Text RAMA_55201_09021382126126_0008118205_0203128701_08_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (223kB) | Request a copy

Abstract

As music genres diversify and online music libraries grow, the need for automated, accurate genre classification has become essential for efficient music organization and recommendation. In this research, we developed a music genre classifier using a custom Convolutional Neural Network (CNN) trained on mel-spectrogram images derived from the GTZAN dataset. The GTZAN dataset is a widely used benchmark in music genre classification and comprises 1,000 music audio samples, each 30 seconds in duration, categorized into 10 distinct genres: blues, classical, country, disco, hiphop, jazz, metal, pop, reggae, and rock. These audio samples were preprocessed by converting them to mel-spectrograms using a base frequency of 22,050 Hz, capturing the spectral characteristics essential for genre differentiation. The dataset was then split into an 80:20 ratio for training and validation. After exploring and testing 20 CNN architectures, 10 models achieved over 50% validation accuracy, with the best model achieves 69.5% accuracy. This work highlights the potential and challenges of designing an effective CNN model specifically for genre classification tasks.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Klasifikasi Genre Musik, Convolutional Neural Network (CNN), Mel-Spectrogram, Dataset GTZAN, Deep Learning.
Subjects:	Q Science > Q Science (General) > Q300-390 Cybernetics > Q325.5 Machine learning Q Science > Q Science (General) > Q334-342 Computer science. Artificial intelligence. Algorithms. Robotics. Automation.
Divisions:	09-Faculty of Computer Science > 55201-Informatics (S1)
Depositing User:	Muhamad Dwirizqy Wimbassa
Date Deposited:	31 Dec 2024 02:58
Last Modified:	31 Dec 2024 02:58
URI:	http://repository.unsri.ac.id/id/eprint/161995

Actions (login required)

View Item