Text
Analisis Citra CT Scan Kanker Paru Berdasarkan Ciri Teklstur Gray level Co-Occurrence Matrix (GLCM) Dan ciri Morfologi Menggunakan Jaringan Syaraf Tiruan Propagasi balik 006,4
INTISARI
Telah dilakukan penelitian mengenai analisis citra CT Scan kanker paru berdasarkan ciri
tekstur Gray Level Co-occurrence Matrix (GLCM) dan ciri morfologi menggunakan
jaringan syaraf tiruan propagasi balik. Kanker paru merupakan kanker yang paling umum
terjadi didunia. Pada tahun 2012, terdapat 1,8 juta kasus baru dan 1,6 juta kematian akibat
kanker. Penelitian ini bertujuan menganalisa citra CT Scan kanker paru berdasarkan ciri
tekstur Gray Level Co-occurrence Matrix (GLCM) dan ciri morfologi menggunakan
jaringan syaraf tiruan propagasi balik serta menghitung akurasi pengujian jaringan syaraf
tiruan propagasi balik. Penelitian dilakukan melalui tahapan segmentasi thresholding,
ekstraksi ciri dan klasifikasi. Ekstraksi ciri tekstur dan morfologi diperoleh dari segmentasi
thresholding. Hasil ekstraksi ciri berupa nilai kontras, korelasi, energi, homogenitas dan
area ratio kemudian digunakan sebagai masukan pada proses pelatihan dan pengujian
jaringan syaraf tiruan propagasi balik. Proses pelatihan dilakukan selama 4 sekon dengan
jumlah iterasi sebanyak 113 kali. Pada proses pelatihan dari 86 citra data latih, 85
terklasifikasi dengan baik sehingga diperoleh akurasi mencapai 98,83% dan pada pengujian
dari 57 citra data uji, 56 citra terklasifikasi dengan benar dan diperoleh akurasi pengujian
mencapai 98,24%.
Kata kunci : Kanker paru, citra CT Scan, ciri tekstur GLCM, ciri morfologi, jaringan
syaraf tiruan.
ABSTRACT
The research about analysis of CT Scan image of lung cancer based on texture feature
Gray Level Co-occurrence Matrix (GLCM) and morphological using neural network back
propagation has been done. Lung cancer is cancer that general occurred in the word. In
2012, 1,8 million new cases lung cancer and 1,6 million mortality because lung cancer.
The research aim to analysis CT Scan image of lung cancer based on texture feature Gray
Level Co-occurrence Matrix (GLCM) and morphology using artificial neural network back
propagation and calculated accuracy of testing artificial neural network back propagation.
This research conducts pass through stages of segmentation, feature extraction and
classfication. Texture and morphology feature extraction are obtained from the
thresholding segmentation. The result of feature extraction are value contrast, correlation,
energy, homogeneity and area ratio then used to input in process training and testing using
neural network back propagation. Process training is conducts since 4 second with number
of iteration 113 iteration. In proces training from 86 train data image, 85 image is able to
classified, so acurracy of training up to 98,83% and in process testing from 57 test data,
56 test data is able to classified, so test accuracy value up to 98,24%.
Keywords: Lung cancer, CT Scan image, texture feature GLCM, morphological feature,
artificial neural network.
1083D16IV | 006.4 SAI a | Perpustakaan FSM Undip (Referensi) | Tersedia |
Tidak tersedia versi lain