Perbandingan Metode Static thresholding dan Multi Otsu Thresholding pada Segmentasi Citra AML M0 dan AML 1

Indah Asmari, Ersi (2018) Perbandingan Metode Static thresholding dan Multi Otsu Thresholding pada Segmentasi Citra AML M0 dan AML 1. Other thesis, Universitas Sebelas Maret.

[img] PDF - Published Version
Download (248Kb)

    Abstract

    ABSTRAK Proses diagnosa Acute Myeloid Leukemia (AML) dapat dilakukan dengan memanfaatkan teknologi pengolahan citra yang terdiri dari pre-processing, segmentasi, dan ekstraksi ciri. Segmentasi citra digunakan untuk memisahkan objek dari background-nya. Salah satu contoh metode segmentasi yaitu Static Thresholding dan Multi Otsu thresholding. Penelitian ini bertujuan untuk membandingkan metode segmentasi static thresholding dan multi otsu thresholding dalam proses diagnosis awal AML M0 dan M1. Metode pengolahan citra yang digunakan meliputi median filtering, konversi warna YCbCr, static thresholding, multi otsu thresholding, operasi morfologi, sedangkan identifikasi diterapkan menggunakan metode Naïve Bayes Classifier dengan data uji berupa diameter WBC, rasio nukleus, dan putaran nukleus. Proses pengujian menggunakan 29 Citra AML M0 dan 30 citra AML M1. Hasil pengujian menunjukkan multi otsu thresholding menghasilkan akurasi 83.81% sedangkan static thresholding menghasilkan akurasi 75.35%. Berdasarkan akurasinya, segmentasi multi otsu thresholding memberikan hasil klasifikasi yang lebih baik dibandingkan metode static thresholding. Kata Kunci: Acute Myeloid Leukemia (AML), segmentasi citra, multi otsu thresholding, static thresholding ABSTRACT The process of diagnosing Acute Myeloid Leukemia (AML) can be done by utilizing image processing techniques consisting of pre-processing, segmentation, and feature extraction. Image segmentation is used to separate objects from the background. One example of the segmentation method is static thresholding and multi otsu threshholding. This study aims to compare segmentation methods between static thresholding and multi otsu thresholding in the process of AML M0 and M1 as the initial diagnosis. Image processing methods that are used in this research are YCbCr color space, median filtering, static thresholding, multi otsu thresholding, morphological operation. The process of identificating cell uses Naïve Bayes Classifier with test data in the form of WBC diameter, ratio of nukleus, and roundness of nukleus. Data to test process are 29 AML M1 images and 30 AML M2 images. It was shown that segmentation with multi otsu thresholding method give a better classification result with an accuracy of 83.81% whereas on static thresholding method only 75.34%. Keywords: Acute Myeloid Leukemia (AML), image segmentation, multi otsu thresholding, static thresholding

    Item Type: Thesis (Other)
    Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
    Z Bibliography. Library Science. Information Resources > Z665 Library Science. Information Science
    Divisions: Fakultas Matematika dan Ilmu Pengetahuan Alam
    Fakultas Matematika dan Ilmu Pengetahuan Alam > Informatika
    Depositing User: Pratama Wisnu Samodro
    Date Deposited: 03 Jul 2018 01:48
    Last Modified: 03 Jul 2018 01:48
    URI: https://eprints.uns.ac.id/id/eprint/41390

    Actions (login required)

    View Item