Segmentasi Citra Kanker Paru-Paru Menggunakan Metode K-Means Clustering Dan Fuzzy C-Means Clustering

Astaghfirul, Liga Persada (2021) Segmentasi Citra Kanker Paru-Paru Menggunakan Metode K-Means Clustering Dan Fuzzy C-Means Clustering. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
01111640000066-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only

Download (1MB) | Request a copy

Abstract

Dengan terus bertambahnya jumlah pasien kanker paru-paru dan terbatasnya tenaga ahli radiologi onkologi, hal ini membuat proses segmentasi kanker paru-paru oleh ahli memakan waktu yang lama. Maka dari itu dibutuhkan metode segmentasi citra kanker paru-paru yang cepat dan andal. Penelitian ini bertujuan untuk melakukan perbandingan antara dua metode segmentasi yaitu metode k-means clustering dan fuzzy c-means clustering. Data yang digunakan adalah 10 citra dari 22 citra CT scan pada data non-small cell lung cancer (NSLC)-Radiomics-Interobserver1, yang didapatkan dari The Cancer Imaging Archive (TCIA). Pengolahan citra dilakukan menggunakan software Matlab 2020a. Sebelum dilakukannya segmentasi, dilakukan pre-processing dengan menggunakan median filter dan contrast stretching. Setelah disegmentasi dan diekstraksi, citra dibandingkan dengan citra ground truth yang telah disegmentasi oleh ahli radiologi onkologi. Hasil penelitian menunjukkan bahwa metode k-means clustering lebih baik dibandingkan metode fuzzy c-means dengan nilai presisi dan F1-score berturut-turut adalah 91,22% dan 84,05% serta 91,22% dan 76,62%. ================================================================================================ With the increasing number of lung cancer patients and the limited number of radiological oncologists experts, segmenting lung cancer by experts take a long time. Therefore, a fast and reliable image segmentation method is needed. This study aims to make a comparison between two segmentation methods, namely the k-means clustering method and the fuzzy c-means clustering method. The data used were 10 images from 22 CT scan images on non-small cell lung cancer (NSLC) -Radiomics-Interobserver1 data obtained from The Cancer Imaging Archive (TCIA). Image processing was carried out using Matlab 2020a. Before segmentation, pre-processing was carried out using a median filter and contrast stretching. After being segmented and extracted, the image was compared with the ground truth image that has been segmented done by oncology radiologist. The results showed that the k-means clustering method was better than the fuzzy c-means method with the precision and F1-scores of 91.22%, 84.05%, and 91.22%, 76.62%, respectively.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Citra paru-paru, Fuzzy c-means clustering, K-means clustering, Segmentasi, Fuzzy c-means clustering, Image processing, K-means clustering, Segmentation
Subjects: Q Science > QC Physics
Divisions: Faculty of Science and Data Analytics (SCIENTICS) > Physics > 45201-(S1) Undergraduate Thesis
Depositing User: Liga Persada Astaghfirul
Date Deposited: 07 Mar 2021 09:42
Last Modified: 07 Mar 2021 09:42
URI: https://repository.its.ac.id/id/eprint/83694

Actions (login required)

View Item View Item