Rancang Bangun Aplikasi Identifikasi Suara Berbasis Web Menggunakan Convolutional Neural Network

Dzaka, Muhammad Abyan (2023) Rancang Bangun Aplikasi Identifikasi Suara Berbasis Web Menggunakan Convolutional Neural Network. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 05111640007003-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (6MB) | Request a copy

Abstract

Setiap manusia memiliki karakteristik unik dari fisiologis dan perilaku yang dapat menjadi pengenalan identitas diri. Pengenalan identitas seseorang menggunakan karakteristik unik pada manusia biasa dikenal dengan pengenalan biometrik. Ada berbagai pengenalan biometrik seperti sidik jari, wajah, iris mata, dan pengenalan suara. Pada pengenalan biometrik suara, memiliki keuntungan tidak memerlukan biaya yang besar karena tidak memerlukan perangkat khusus selain microphone dan ukuran file-nya yang kecil. Pemilihan ekstraksi fitur yang baik merupakan salah satu faktor yang akan mempengaruhi hasil akurasi identifikasi suara. Mel-Frequency Cepstral Coefficient (MFCC) dipilih sebagai metode untuk ekstraksi fitur sebab MFCC dikenal memiliki akurasi yang tinggi. Selain itu, metode klasifikasi Convolutional Neural Network (CNN) juga telah diterapkan untuk identifikasi suara iringan serta nyanyian dan terbukti lebih baik daripada metode Multi-Layer Perceptron lainnya. Oleh karena itu, dengan terciptanya aplikasi identifikasi suara berbasis web diharapkan mampu mengidentifikasi pemilik suara sehingga dapat digunakan sebagai salah satu opsi untuk keamanan biometrik.
==================================================================================================================================
Every human being has unique physiological and behavioral characteristics that can become self-identity recognition. The identification of a person's identity using unique characteristics in ordinary humans is known as biometric recognition. There are various biometric recognition such as fingerprint, face, iris and voice recognition. In voice biometric recognition, it has the advantage of not requiring a large amount of money because it does not require special devices other than a microphone and small file size. Selection of a good feature extraction is one of the factors that will affect the results of the accuracy of voice identification. Mel Frequency Cepstral Coefficient (MFCC) was chosen as a method for feature extraction because MFCC is known to have high accuracy. In addition, the Convolutional Neural Network (CNN) classification method has also been applied to identify accompaniment and singing sounds and is proven to be better than other Multi-Layer Perceptron methods. Therefore, the creation of a web-based voice identification application is expected to be able to identify the owner of the voice so that it can be used as an option for biometric security.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Identifikasi, Convolutional Neural Network, Mel-Frequency Cepstral Coefficient, Suara, Web. Identification, Convolutional Neural Network, Mel-Frequency Cepstral Coefficient, Voice, Web.
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Q Science > QA Mathematics > QA76.754 Software architecture. Computer software T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK105.8883 Web authoring software (include web server) T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.888 Web sites--Design. Web site development.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Muhammad Abyan Dzaka
Date Deposited:	02 Aug 2023 05:00
Last Modified:	02 Aug 2023 05:00
URI:	http://repository.its.ac.id/id/eprint/101632

Actions (login required)

View Item