Klasifikasi Suara Batuk COVID-19 Berbasis Deep Learning.

Zanjabila, Zanjabila (2022) Klasifikasi Suara Batuk COVID-19 Berbasis Deep Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 02311840000041-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (4MB)

Abstract

Pandemi COVID-19 merupakan permasalahan umum negara-negara di dunia. Salah satu pencegahan penularan dengan diagnosis awal, tetapi diagnosis COVID-19 memiliki kekurangan. Kekurangan tersebut adalah akurasi yang berkurang seiring berjalannya waktu dari penderita. Oleh karena itu, dikembangkan metode diagnosis COVID-19 berdasarkan suara batuk yang merupakan gejala paling umum pada penderita COVID-19. Penelitian ini bertujuan untuk mengetahui performa deep learning dengan menggunakan pre-processing block untuk klasifikasi suara batuk COVID-19 berdasarkan metrik UA (unweighted accuracy) dengan menggabungkan beberapa dataset. Selain itu, ingin mengetahui ukuran fail model dan real-time factor (RTF) yang didapat dari deep learning untuk klasifikasi suara batuk COVID-19. Dalam penelitian ini dilakukan beberapa tahap pre-processing blocks seperti deteksi batuk dan segmentasi. Hasil penelitian didapatkan unweighted accuracy (UA) sebesar 88,19% dan memiliki ukuran fail model sebesar 938 Mb dengan RTF terbaik sebesar 0,31 sehingga inference program bisa dikatakan real-time pada jenis komputer yang digunakan.===================================================================================================================================
The COVID-19 pandemic is a common problem in countries around the world. One of the prevention of transmissions with early diagnosis, but the diagnosis of COVID-19 has shortcomings. The most important shortcoming is the duration of test and its performance. Therefore, a COVID-19 diagnostic method was developed based on the sound of coughing which is the most common symptom in COVID-19 sufferers. This study aims to determine the performance of deep learning by using a pre-processing block for the classification of COVID-19 cough sounds based on the UA (unweighted accuracy) metric by combining several datasets. In addition, this research want to know the model file size and real-time factor (RTF) obtained from deep learning for the classification of COVID-19 cough sounds. In this study, several pre-processing blocks were carried out such as cough detection and segmentation. The results of the study obtained unweighted accuracy (UA) of 88.19% and has a model file size of 938 Mb with the best RTF of 0.31. The inference program can be said to be real-time for the computer used for the experiments.

Item Type:	Thesis (Other)
Additional Information:	RSF 006.31 Zan k-1 2022
Uncontrolled Keywords:	Klasifikasi, Batuk, COVID-19, Deep learning, Deteksi Batuk, Segmentasi, Augmentasi. Classification, Cough, COVID-19, Deep Learning, Cough Detection, Segmentation, Augmentation.
Subjects:	R Medicine > R Medicine (General) > R858 Deep Learning
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Mr. Marsudiyana -
Date Deposited:	11 May 2026 03:54
Last Modified:	11 May 2026 03:54
URI:	http://repository.its.ac.id/id/eprint/133113

Actions (login required)

View Item