Penerapan Metode Mel Frequency Ceptral Coefficient dan Learning Vector Quantization untuk Text-Dependent Speaker Identification
DOI:
https://doi.org/10.61769/telematika.v11i1.147Keywords:
Pengenalan suara, Learning Vector Quantization, Mel Frequency Ceptral Coefficients, ekstraksi fitur suara, seleksi fitur, Text-Dependent.Abstract
Layanan keamanan pada umumnya menggunakan kata sandi untuk membatasi dan mengontrol akses layanan tersebut.Kata sandi yang biasa digunakan sering kali berbentuk teks.Penggunaan kata sandi dengan bentuk teks dianggap masih kurang aman karena sering kali terjadi kebocoran. Maka dari itu dibutuhkanlah bentuk lain dari kata sandi, untuk meningkatkan keamanan dalam mengakses layanan atau data tertentu. Salah satunya adalah dalam bentuk suara. Sistem ini berbasis pada input berupa file audio dengan data ucapan yang bergantung pada teks atau text-dependent dengan output adalah identitas pembicara yang teridentifikasi. Pada penelitian ini, sistem pengenalan pembicara dibuat untuk dapat mengenali suara pembicara dengan menggunakan Mel-Frequency Cepstral Coefficients yang digunakan untuk melakukan ekstraksi fitur dari data suara sehingga dihasilkan fitur-fitur yang mewakili pembicara tersebut dan metode Learning Vector Quantization yang digunakan untuk melatih data-data hasil ekstraksi dan mencocokan data latih dengan data baru sehingga didapatkan identitas dari pembicara berdasarkan suara tersebut. Dari hasil pengujian pada sistem ini, didapatkan identification rate tertinggi adalah 88.9% dengan menggunakan data dengan durasi sekitar 8 detik.
Security services generally use a password to restrict and control access to its services. Many password used is often in the text form. This type of password is considered less secure because it can be obtained by unauthorized people. Other forms of password are required to increase the security in accessing services or specific data such as voices. This system is based on the input of an audio file such as utterance that depends on text or text-dependent. In this study, the speaker recognition system is made to recognize the speaker of an audio file using Mel-Frequency Ceptral Coefficients for extracting voice data to produce features that represent the speaker and Learning Vector Quantization (LVQ) to train the data extraction and matching training data with new data to obtain the identity of the speaker based on the sound. From the experiment result, obtained the highest identification rate is 88.9% using data with a duration about 8 seconds.
References
Kshamamayee Dash, Debananda Padhi, Bhoomika Panda, and Sanghamitra Mohanty, "Speaker Identification Using Mel Frequency Cepstral Coefficient And Bpnn," International Journal of Advanced Research in Computer Science and Software Engineering, vol. 2, no. 4, April 2012.
Zhizheng Wu, Anthony Larcher, and Kong Aik Lee, "Vulnerability evaluation of speaker verification under voice conversion spoofing: the effect of text constraints.," in INTERSPEECH, 2013, pp. 950-954.
Utpal Bhattacharjee, "A Comparative Study Of LPCC And MFCC Features For The Recognition Of Assamese Phonemes," International Journal of Engineering Research and Technology, vol. 2, no. 1, January 2013.
Penghua LI, Shunxing Zhang, Huizong Feng, and Yuanyuan Li, "Speaker Identification Using Spectrogram And Learning Vector Quantization," Journal of Computational Information Systems, vol. 11, no. 9, 2015.
Geeta Nijhawan and M.K Soni, "Speaker Recognition Using Mfcc And Vector Quantisation," International Journal on Recent Trends in Engineering and Technology, vol. 11, no. 1, Juli 2014.
Richard G Lyons, Understanding Digital Signal Processing 3rd Edition. Boston: Prentice Hall, 2011.
Laurene Fausett, Fundamental of Neural Networks: Architectures, Algorithms, and Applications.: Prentice Hall, 1994.
Downloads
Published
Issue
Section
License
You are free to:
- Share — copy and redistribute the material in any medium or format for any purpose, even commercially.
- Adapt — remix, transform, and build upon the material for any purpose, even commercially.
- The licensor cannot revoke these freedoms as long as you follow the license terms.
Under the following terms:
- Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
- ShareAlike — If you remix, transform, or build upon the material, you must distribute your contributions under the same license as the original.
- No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.
Notices:
You do not have to comply with the license for elements of the material in the public domain or where your use is permitted by an applicable exception or limitation.
No warranties are given. The license may not give you all of the permissions necessary for your intended use. For example, other rights such as publicity, privacy, or moral rights may limit how you use the material.