ALMAHMUD, DERA (2023) PENGENALAN EKSPRESI WAJAH PENGUCAPAN HURUF VOKAL MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK. Undergraduate thesis, Universitas Stikubank.
![[thumbnail of HALAMAN JUDUL]](/style/images/fileicons/application_pdf.png) |
PDF (HALAMAN JUDUL)
Download (1MB) |
|
PDF (ABSTRAK)
Download (233kB) |
|
PDF (BAB I)
Download (223kB) |
|
PDF (BAB II)
Restricted to Repository staff only Download (1MB) |
|
PDF (BAB III)
Restricted to Repository staff only Download (1MB) |
|
PDF (BAB IV)
Restricted to Repository staff only Download (448kB) |
|
PDF (BAB V)
Restricted to Repository staff only Download (219kB) |
|
PDF (DAFTAR PUSTAKA)
Download (439kB) |
|
PDF (LAMPIRAN)
Restricted to Repository staff only Download (2MB) |
Abstract
Ekspresi merupakan salah satu cara manusia untuk melakukan komunikasi antara satu dengan yang lain yang dilakukan secara sadar maupun tidak. Ekspresi dari seorang manusia dapat menunjukkan tujuan dan maksud dari seorang manusia. Dengan berkembangnya teknologi machine learning, membaca ekspresi seseorang dapat dilakukan secara otomatis. Tujuan penelitian ini adalah untuk membuat suatu model Convolutional Neural Network (CNN) yang mampu melakukan klasifikasi ekspresi pengucapan huruf vokal. Data yang digunakan pada penelitian ini didapat dari relawan dengan mengambil foto ekspresi pengucapan huruf vokal dan memperoleh 300 data. Hasil penelitian dengan jumlah 75 epoch menunjukkan tingkat akurasi sebesar 90% dengan memanfaatkan fine-tuning pada pre-trained model. Penggunaan jumlah epoch pada pelatihan mempengaruhi kemampuan pada model. Selain itu proses augmentasi data juga dapat dimanfaatkan untuk mengatasi kekurangan data dan membantu meningkatkan kemampuan model. Penentuan jumlah epoch dapat melalui pengamatan tingkat accuracy, loss, val accuracy dan val loss. Expression is one of the ways humans communicate with each other, consciously or unconsciously. An individual's expression can indicate their intentions and purposes. With the development of machine learning technology, reading someone's expression can now be done automatically. The goal of this research is to create a Convolutional Neural Network (CNN) model that can classify expressions of vowel pronunciation. The data used in this study was obtained from volunteers by taking photos of vowel pronunciation expressions and obtaining 300 data. The results of the study with 75 epochs showed an accuracy rate of 90% by utilizing fine-tuning on a pre-trained model. The number of epochs used in training affects the ability of the model. In addition, data augmentation can also be used to overcome data deficiencies and help improve the model's performance. The determination of the number of epochs can be based on observations of the accuracy rate, loss, val accuracy, and val loss.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | SKR.I.05.01.2250 NIM.18.01.53.0098 |
| Uncontrolled Keywords: | Convolutional Neural Network, Face Expression Recognition, Augmentasi, Fine-Tuning. |
| Subjects: | Q Science > QA Mathematics Q Science > QA Mathematics > QA75 Electronic computers. Computer science |
| Faculty / Institution: | Fakultas Teknologi Informasi > Program Studi Teknik Informatika |
| Depositing User: | Teteh Hayati |
| Date Deposited: | 17 May 2023 02:27 |
| Last Modified: | 17 May 2023 02:27 |
| URI: | https://eprints.unisbank.ac.id/id/eprint/9415 |
Actions (login required)
 |
View Item |
