Please use this identifier to cite or link to this item: https://repositori.uma.ac.id/handle/123456789/17072
Title: Perbandingan Arsitektur Resnet-50 dan Inceptionv3 dalam Klasifikasi Covid 19 Berdasarkan Citra X Ray
Other Titles: Comparison of Resnet-50 and Inceptionv3 Architecture in Classification of Covid 19 Based on X Ray Image
Authors: Ryandra, Muhammad Farhan Dwi
metadata.dc.contributor.advisor: Muhathir
Syah, Rahmad
Keywords: covid;resnet 50;inception v3
Issue Date: 11-Jun-2022
Publisher: Universitas Medan Area
Series/Report no.: NPM;178160047
Abstract: Penyakit yang akhir-akhir ini menjadi perbincangan hangat masyarakat di seluruh dunia ditemukan pertama kali pada kota Wuhan, Provinsi Hubei, China yang kemudian dilaporkan ke Organisasi Kesehatan Dunia yaitu WHO(World Health Organization) pada 31 Desember 2019 merupakan wabah penyakit yang menginfeksi saluran pernapasan yang disebabkan oleh virus yang disebut dengan istilah Severe Acute Respiratory Syndrome Corona Virus 2 (SARS-CoV-2). Convolutional Neural Network (CNN) dapat dimanfaatkan untuk mendeteksi paru-paru terinfeksi COVID-19 melalui citra x-ray. Model deep learning yang digunakan pada penelitian ini adalah arsitektur resnet-50 dan inception v3. Penelitian ini menggunakan beberapa parameter yaitu epoch, batch size dan optimizer. Hasil penelitian menunjukkan bahwa arsitektur resnet-50 maupun inception v-3 mempunyai epoch yang paling baik adalah epoch yang berjumlah 25, dengan batch yang paling baik adalah 200 dan optimizer yang paling baik adalah adam. Secara keseluruhan hyperparameter yang paling optimal pada kondisi epoch 25, batch 200 optimizer adam dengan akurasi berjumlah 99% dimana waktu komputasi dimana inception v-3 membutuhkan total waktu 6 jam sedangkan resnet-50 9 jam 21 menit. Arsitektur inception v3 adalah arsitektur yang paling optimal dalam klasifikasi covid-19 berdasarkan citra x-ray. The disease, which has recently become a hot topic of conversation among people around the world, was first discovered in the city of Wuhan, Hubei Province, China, which was later reported to the World Health Organization, namely the WHO (World Health Organization) on December 31, 2019 as an outbreak of a disease that infects the respiratory tract. It is caused by a virus known as Severe Acute Respiratory Syndrome Corona Virus 2 (SARS-CoV-2). Convolutional Neural Network (CNN) can be used to detect COVID-19-infected lungs through x-ray images. The deep learning model used in this research is the resnet-50 architecture and inception v3. This study uses several parameters, namely epoch, batch size and optimizer. The results showed that the resnet-50 and inception v-3 architectures had the best epochs of 25, with the best batch being 200 and the best optimizer being adam. Overall, the most optimal hyperparameter is in epoch 25, batch 200 optimizer adam with 99% accuracy where the computation time where Inception v-3 takes a total of 6 hours while resnet-50 is 9 hours 21 minutes. Inception v-3 architecture is the most optimal architecture in the classification of covid-19 based on x-ray images.
Description: 39 Halaman
URI: https://repositori.uma.ac.id/handle/123456789/17072
Appears in Collections:SP - Informatic Engineering

Files in This Item:
File Description SizeFormat 
178160047 - Muhammad Farhan Dwi Ryandra - Fulltext.pdfCover, Abstract, Chapter I, II, III, V, Bibliography1.12 MBAdobe PDFView/Open
178160047 - Muhammad Farhan Dwi Ryandra - Chapter IV.pdf
  Restricted Access
Chapter IV960.39 kBAdobe PDFView/Open Request a copy


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.