Azis, Aaz Muhammad Hafidz (2021) Implementasi algoritma CNN untuk mendeteksi pelafalan huruf hijaiyah berharakat kasrah. Diploma thesis, UIN Sunan Gunung Djati Bandung.
|
Text (COVER)
1_cover.pdf Download (239kB) | Preview |
|
|
Text (ABSTRAK)
2_Abstrak.pdf Download (141kB) | Preview |
|
|
Text (DAFTAR ISI)
3_Daftar Isi.pdf Download (694kB) | Preview |
|
|
Text (BAB I)
4_BAB 1.pdf Download (778kB) | Preview |
|
![[img]](https://etheses.uinsgd.ac.id/style/images/fileicons/text.png) |
Text (BAB II)
5_BAB II.pdf Restricted to Registered users only Download (1MB) | Request a copy |
|
|
Text (BAB III)
6_BAB III.pdf Restricted to Registered users only Download (931kB) | Request a copy |
|
|
Text (BAB IV)
7_BAB IV.pdf Restricted to Registered users only Download (850kB) | Request a copy |
|
|
Text (BAB V)
8_BAB V.pdf Restricted to Registered users only Download (385kB) | Request a copy |
|
|
Text (DAFTAR PUSTAKA)
9_Daftar Pustaka.pdf Restricted to Registered users only Download (454kB) | Request a copy |
Abstract
INDONESIA: Sebagai seorang muslim yang baik dianjurkan untuk senantiasa membaca al-Quran dengan benar setidaknya sepuluh ayat setiap hari. Al-Qur’an disusun dengan huruf hijaiyah dengan makhraj yang berbeda. Selain itu terdapat harakat yang digunakan untuk mempermudah membaca huruf arab. Namun tidak semua orang islam mampu melafalkan bacaan al-Qur’an dengan benar. Penelitian ini mengusulkan sebuah metode yang mampu mengenalkan dan mengajarkan lafal huruf hijaiyah menggunakan teknologi speech recognition yang dapat mengenali suara baik per kata maupun per kalimat. Suara kemudian diekstraksi menggunakan model Mel- frequency cepstral coefficients (MFCC) dan selanjutnya diklasifikasikan menggunakan model deep learning dengan algoritma CNN. Penelitian ini berhasil mengklasifikasikan harakat kasrah menggunakan CNN dengan metode MFCC dengan kinerja terbaik pada akurasi 62.45%, presisi 75%, recall 50% dan f1-score sebesar 58%. ENGLISH: As a good Muslim it is recommended to always read Al-Qur'an correctly at least ten verses every day. Al-Qur'an is arranged in hijaiyah letters with different makhraj. In addition there is a vowel that is used to make it easier to read Arabic letters. However, not all Muslims are able to pronounce the recitation of Al-Qur'an correctly. This study proposes a method that is able to introduce and teach hijaiyah pronunciation using speech recognition technology that can recognize sounds both words and sentences. The sound is then extracted using the Mel-frequency cepstral coefficients (MFCC) model and then classified using a deep learning model with the CNN algorithm. This study has succeeded in classifying the harakat kasrah using CNN with the MFCC method with the best performance at an accuracy of 62.45%, 75% precision, 50% recall and an f1-score of 58%.
| Item Type: | Thesis (Diploma) |
|---|---|
| Uncontrolled Keywords: | Hijaiyah; Kasrah; CNN; MFCC; Speech Recognition |
| Subjects: | Al-Qur'an (Al Qur'an, Alquran, Quran) dan Ilmu yang Berkaitan Educational Institutions, Schools and Their Activities Engineering |
| Divisions: | Fakultas Sains dan Teknologi > Program Studi Teknik Informatika |
| Depositing User: | Aaz Muhammad Hafidz Azis |
| Date Deposited: | 27 Jul 2021 03:26 |
| Last Modified: | 27 Jul 2021 03:26 |
| URI: | https://etheses.uinsgd.ac.id/id/eprint/40935 |
Actions (login required)
 |
View Item |
