• Login
    View Item 
    •   USU-IR Home
    • Faculty of Computer Science and Information Technology
    • Department of Information Technology
    • Undergraduate Theses
    • View Item
    •   USU-IR Home
    • Faculty of Computer Science and Information Technology
    • Department of Information Technology
    • Undergraduate Theses
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Klasifikasi Penyakit Pneumonia (Paru-Paru Basah) Menggunakan Probabilistic Neural Network

    View/Open
    Fulltext (3.593Mb)
    Date
    2021
    Author
    Manurung, Agnes
    Advisor(s)
    Andayani, Ulfi
    Jaya, Ivan
    Metadata
    Show full item record
    Abstract
    Pneumonia is an infection of the lung tissue (alveoli) and the air sacs are filled with pus or fluid caused by bacteria, viruses, or fungi. Pneumonia is often found in outpatients, elderly people and causes a high number of deaths among children under five all over the world. Pneumonia is generally diagnosed manually by the doctor through analysis of Chest X-Ray (CXR) or X-Ray image, so it takes time. To overcome this problem, a method is needed to classify pneumonia through CXR images automatically. The purpose of this research is to facilitate doctors and medical workers in classifying pneumonia. The method proposed in this research is Probabilistic Neural Network to classify normal, bronchopneumonia and lobar. The stages before classification are preprocessing (scaling, grayscaling, contrast), segmentation (thresholding, canny detection) and feature extraction using Moment Invariants. This research used 323 training data and 60 testing data that taken from data.mendeley.com, radiopaedia.org and www.kaggle.com. After the testing was done, it was concluded that the proposed method was able to classify pneumonia with 95% accuracy, 95% precision and 95% recall. Misclassification is influenced by the similarity between the image categories and the neck and arms contained in the image are considered by the system as part of the image.
     
    Pneumonia merupakan infeksi jaringan paru-paru (alveoli) di mana kantung udara dipenuhi nanah atau cairan yang disebabkan oleh bakteri, virus, maupun jamur. Pneumonia sering dijumpai pada pasien rawat jalan, umur lanjut usia dan menyebabkan jumlah kematian yang tinggi pada balita di seluruh dunia. Pneumonia umumnya didiagnosis secara manual oleh dokter melalui analisis citra Chest X-Ray (CXR), sehingga diperlukan waktu yang lebih lama. Untuk mengatasi masalah tersebut, maka dibutuhkan metode untuk mengklasifikasi pneumonia melalui citra CXR secara otomatis. Tujuan penelitian ini untuk mempermudah dokter dan pekerja medis dalam mengklasifikasikan pneumonia. Metode yang diajukan pada penelitian ini adalah Probabilistic Neural Network untuk mengklasifikasi citra normal, bronchopneumonia dan lobar. Tahapan yang dilakukan sebelum klasifikasi adalah preprocessing (scaling, grayscaling, contrast), segmentation (thresholding, canny detection) dan feature extraction menggunakan Moment Invariants. Penelitian ini menggunakan 323 data training dan 60 data testing yang diambil dari data.mendeley.com, radiopaedia.org dan www.kaggle.com. Setelah pengujian dilakukan, diperoleh kesimpulan bahwa metode yang diajukan mampu melakukan klasifikasi pneumonia dengan akurasi sebesar 95%, precision 95% dan recall 95%. Kesalahan klasifikasi dipengaruhi oleh kemiripan antara kategori citra dan bagian leher serta lengan yang terdapat pada citra dianggap sistem sebagai bagian dari citra.

    URI
    http://repositori.usu.ac.id/handle/123456789/44250
    Collections
    • Undergraduate Theses [800]

    Repositori Institusi Universitas Sumatera Utara (RI-USU)
    Universitas Sumatera Utara | Perpustakaan | Resource Guide | Katalog Perpustakaan
    DSpace software copyright © 2002-2016  DuraSpace
    Contact Us | Send Feedback
    Theme by 
    Atmire NV
     

     

    Browse

    All of USU-IRCommunities & CollectionsBy Issue DateTitlesAuthorsAdvisorsKeywordsTypesBy Submit DateThis CollectionBy Issue DateTitlesAuthorsAdvisorsKeywordsTypesBy Submit Date

    My Account

    LoginRegister

    Repositori Institusi Universitas Sumatera Utara (RI-USU)
    Universitas Sumatera Utara | Perpustakaan | Resource Guide | Katalog Perpustakaan
    DSpace software copyright © 2002-2016  DuraSpace
    Contact Us | Send Feedback
    Theme by 
    Atmire NV