Model Deteksi COVID-19 dari Citra CT Scan Dada Menggunakan DenseNet-121

Abdillah Hanif Maulana, Ir. Nazrul Effendy, S.T., M.T., Ph.D., IPM.; Ir. Nopriadi, S.T., M.Sc., Ph.D., IPM.

2024 | Skripsi | FISIKA TEKNIK

Diagnosis primer COVID-19 adalah uji RT-PCR, namun ditemukan bahwa RT-PCR memiliki kelemahan sensitivitas yang rendah pada fase awal infeksi. CT scan dada memiliki kelebihan sensitivitas yang lebih tinggi pada fase awal infeksi dibanding RT-PCR, sehingga dapat digunakan sebagai komplemen dari uji RT-PCR untuk membantu mengurangi penyebaran COVID-19 akibat hasil false negative. Untuk membantu tenaga medis, Deep Learning dapat digunakan untuk automasi proses deteksi COVID-19 via citra CT scan dada.

Pada penelitian ini dibangun model deteksi COVID-19 dengan transfer learning DenseNet-121. Dilakukan beberapa variasi yaitu tanpa & dengan fine tuning, juga variasi pada Learning Rate (LR) yaitu LR default (0,001) & LR hasil Learning Rate Finder (0,0001). Model ini dilatih menggunakan callback ReduceLROnPlateau & EarlyStopping. Adapun dataset yang digunakan adalah dataset dengan 3 kelas (Normal, Pneumonia, & COVID-19) dari COVIDx CT-2A yang telah melalui proses undersampling & berbagai jenis augmentasi gambar. Performa model dievaluasi menggunakan berbagai metrik evaluasi yaitu akurasi, sensitivitas, presisi, & spesifisitas.

Diperoleh hasil terbaik adalah model dengan fine tuning & LR hasil Learning Rate Finder. Model ini mampu bekerja dengan baik, dengan akurasi sebesar 97,64%; presisi sebesar 96,49%; sensitivitas sebesar 96,43%; & spesifisitas sebesar 98,25%.

The primary diagnosis of COVID-19 is the RT-PCR test, but it was found that RT-PCR has the disadvantage of low sensitivity in the early phase of infection. Chest CT has the advantage of higher sensitivity in the early phase of infection compared to RT-PCR, so it can be used as a complement to the RT-PCR test to help reduce the spread of COVID-19 due to false negative results. To help medical personnel, Deep Learning can be used to automate the COVID-19 detection process via chest CT images.

In this research, a COVID-19 detection model was built by transfer learning of DenseNet-121. Several variations were done, that is without & with fine tuning, also variations on Learning Rate (LR) which was default LR (0.001) & LR obtained from Learning Rate Finder (0.0001). The model was trained using ReduceLROnPlateau & EarlyStopping callbacks. The dataset used was a dataset made of 3 classes (Normal, Pneumonia, & COVID-19) from COVIDx CT-2A which has gone through an undersampling process & various types of image augmentation. The model performance was then evaluated using various evaluation metrics namely accuracy, sensitivity, precision, & specificity.

The best results obtained were from the model with fine tuning & LR obtained from Learning Rate Finder. This model worked well, with an accuracy of 97.64%; precision of 96.49%; sensitivity of 96.43%; & specificity of 98.25%.

Kata Kunci : COVID-19, CT Scan, Deep Learning, DenseNet-121