Speaker Identification Using a Convolutional Neural Network
Speech, a mode of communication between humans and machines, has various applications, including biometric systems for identifying people have access to secure systems. Feature extraction is an important factor in speech recognition with high accuracy. Therefore, we implemented a spectrogram, which...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Ikatan Ahli Informatika Indonesia
2022-02-01
|
| Series: | Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) |
| Subjects: | |
| Online Access: | http://jurnal.iaii.or.id/index.php/RESTI/article/view/3795 |
| _version_ | 1797334253845872640 |
|---|---|
| author | Suci Dwijayanti Alvio Yunita Putri Bhakti Yudho Suprapto |
| author_facet | Suci Dwijayanti Alvio Yunita Putri Bhakti Yudho Suprapto |
| author_sort | Suci Dwijayanti |
| collection | DOAJ |
| description | Speech, a mode of communication between humans and machines, has various applications, including biometric systems for identifying people have access to secure systems. Feature extraction is an important factor in speech recognition with high accuracy. Therefore, we implemented a spectrogram, which is a pictorial representation of speech in terms of raw features, to identify speakers. These features were inputted into a convolutional neural network (CNN), and a CNN-visual geometry group (CNN-VGG) architecture was used to recognize the speakers. We used 780 primary data from 78 speakers, and each speaker uttered a number in Bahasa Indonesia. The proposed architecture, CNN-VGG-f, has a learning rate of 0.001, batch size of 256, and epoch of 100. The results indicate that this architecture can generate a suitable model for speaker identification. A spectrogram was used to determine the best features for identifying the speakers. The proposed method exhibited an accuracy of 98.78%, which is significantly higher than the accuracies of the method involving Mel-frequency cepstral coefficients (MFCCs; 34.62%) and the combination of MFCCs and deltas (26.92%). Overall, CNN-VGG-f with the spectrogram can identify 77 speakers from the samples, validating the usefulness of the combination of spectrograms and CNN in speech recognition applications. |
| first_indexed | 2024-03-08T08:17:03Z |
| format | Article |
| id | doaj.art-1b381347f0c1447f8c9aad51f2f2a8f2 |
| institution | Directory Open Access Journal |
| issn | 2580-0760 |
| language | English |
| last_indexed | 2024-03-08T08:17:03Z |
| publishDate | 2022-02-01 |
| publisher | Ikatan Ahli Informatika Indonesia |
| record_format | Article |
| series | Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) |
| spelling | doaj.art-1b381347f0c1447f8c9aad51f2f2a8f22024-02-02T06:58:52ZengIkatan Ahli Informatika IndonesiaJurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)2580-07602022-02-016114014510.29207/resti.v6i1.37953795Speaker Identification Using a Convolutional Neural NetworkSuci Dwijayanti0Alvio Yunita Putri1Bhakti Yudho Suprapto2Universitas SriwijayaTeknik Elektro Universitas SriwijayaTeknik Elektro Universitas SriwijayaSpeech, a mode of communication between humans and machines, has various applications, including biometric systems for identifying people have access to secure systems. Feature extraction is an important factor in speech recognition with high accuracy. Therefore, we implemented a spectrogram, which is a pictorial representation of speech in terms of raw features, to identify speakers. These features were inputted into a convolutional neural network (CNN), and a CNN-visual geometry group (CNN-VGG) architecture was used to recognize the speakers. We used 780 primary data from 78 speakers, and each speaker uttered a number in Bahasa Indonesia. The proposed architecture, CNN-VGG-f, has a learning rate of 0.001, batch size of 256, and epoch of 100. The results indicate that this architecture can generate a suitable model for speaker identification. A spectrogram was used to determine the best features for identifying the speakers. The proposed method exhibited an accuracy of 98.78%, which is significantly higher than the accuracies of the method involving Mel-frequency cepstral coefficients (MFCCs; 34.62%) and the combination of MFCCs and deltas (26.92%). Overall, CNN-VGG-f with the spectrogram can identify 77 speakers from the samples, validating the usefulness of the combination of spectrograms and CNN in speech recognition applications.http://jurnal.iaii.or.id/index.php/RESTI/article/view/3795speaker identificationcnnspectrogramfeature extraction |
| spellingShingle | Suci Dwijayanti Alvio Yunita Putri Bhakti Yudho Suprapto Speaker Identification Using a Convolutional Neural Network Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) speaker identification cnn spectrogram feature extraction |
| title | Speaker Identification Using a Convolutional Neural Network |
| title_full | Speaker Identification Using a Convolutional Neural Network |
| title_fullStr | Speaker Identification Using a Convolutional Neural Network |
| title_full_unstemmed | Speaker Identification Using a Convolutional Neural Network |
| title_short | Speaker Identification Using a Convolutional Neural Network |
| title_sort | speaker identification using a convolutional neural network |
| topic | speaker identification cnn spectrogram feature extraction |
| url | http://jurnal.iaii.or.id/index.php/RESTI/article/view/3795 |
| work_keys_str_mv | AT sucidwijayanti speakeridentificationusingaconvolutionalneuralnetwork AT alvioyunitaputri speakeridentificationusingaconvolutionalneuralnetwork AT bhaktiyudhosuprapto speakeridentificationusingaconvolutionalneuralnetwork |