Klasifikasi Massa pada Citra Mammogram Berdasarkan Gray Level Cooccurence Matrix (GLCM)
Refta Listia(1*), Agus Harjoko(2)
(1) 
(2) Jurusan Ilmu Komputer dan Elektronika (JIKE), Fakultas Matematika dan Ilmu Pengetahuan Alam, Universitas Gadjah Mada, Sekip Utara, Bulaksumur, Yogyakarta
(*) Corresponding Author
Abstract
Kanker payudara adalah penyakit yang paling umum dideritaoleh wanitapadabanyak negara. Pemeriksaan kanker payudara dapat dilakukan dengan menggunakan mamografi.Padapenelitianini, pendekatan yang diusulkan bertujuanuntuk mengklasifikasi mammogram berdasarkan tiga kelas yaitukelas normal, tumor jinak, dan tumor ganas. Sistem yang diusulkan terdiri dari empat langkah utamayaitu preprosesing, segmentasi, ekstraksi fitur dan klasifikasi. Padatahappreprosesingakandilakukangrayscale, interpolasi, amoeba mean filter dan segmentasi. Ekstraksi ciri menggunakan Gray Level Cooccurence Matrix (GLCM) danakan dihitung ciri-ciristatistikpada 4 arah (d=1 dan d=2) , GLCM 8 arah(d=1) dan GLCM 16 arah (d=2).Fitur yang digunakanada 5 yaitukontras, energi, entropi, korelasi dan homogenitas. Langkah terakhir adalah klasifikasi menggunakan Backpropagation. Beberapa parameter penting divariasikan dalam proses ini seperti learning rate dan jumlah node dalam lapisan tersembunyi. Hasil penelitian menunjukkan bahwa fitur ekstraksi GLCM 4 arah(denganjarak d=1memiliki akurasi terbaik dalammengklasifikasimammogram yaitusebesar 81,1% dankhususpadaarah akurasi klasifikasidiperolehsebesar 100%.
Abstract
Breast cancer is the most common disease in women in many countries. Breast cancer can be performed using mammography. In this work, an approach is proposed to classify mammogram based on three classes such as normal, benign, and malignant. The proposed system consist of four major steps : preprocessing, segmentation, feature extraction and classification. In preprocessing grayscale, interpolation, amoeba mean filter and segmentation are applicated. Feature extraction using Gray level Cooccurence Matrix (GLCM) and the features will be calculated in 4 angles (d=1 and d= 2), GLCM 8 angles and GLCM 16 angles. The 5 features are contrast, energy, entropy, correlation and homogeneity. The final step is classification using Backpropagation. Some of important parameters will be variated in this process such as learning rate and the number of node in hidden layer. The research result suggest that extraction feature in 4 angles ( and d=1 is the best accuracy for classifying mammogram based on classes 81,1% and especially in accuracy is 100%.
Keywords
Full Text:
PDFReferences
[1] Sjamsuhidajat, R. dan Jong, W., 1997, Buku ajar Ilmu Bedah. Kedokteran EGC, Jakarta.
[2] Putra, D., 2009, Pengolahan Citra Digital, Penerbit Andi, Yogyakarta.
[3] Wibawanto, H., 2011, Analisis Tekstur untuk Diskriminasi Massa Kistik dan Non Kistik pada Citra Ultrasonografi. Disertasi. Fakultas Ilmu Teknik UGM.
[4] Singh, S., Guptaz, P. R., 2011, Breast Cancer detection and Classification using Neural Network, International Journal Of Advanced Engineering Sciences And Technologies. Vol No.6 Issue No. 1, 004 – 009.
[5] Nithya,R. and Santi, B., 2011, Comparative Study on Feature Extraction Method for Breast Cancer Classification, J. Theoritical and Applied Information Technology, Vol. 33 No.2.
[6] Lerallut, R., et al. 2005. Image Filtering Using Morphological Amoebas. Centre de Morphologie Mathématique, École des Mines de Paris.
[7] Puspitaningrum, D. 2006, Pengantar Jaringan Saraf Tiruan. Penerbit Andi. Yogyakarta.
[8] Siang, J.J, 2004, Jaringan Syaraf Tiruan dan Pemrogramannya. Penerbit Andi. Yogyakarta.
[9] Gonzales, R. C., Woods, R. E., 2008, Digital Image Processing, 3rd ed, Prentice Hall, Upper Sadle River, New Jersey, USA
[10] Garcia, M., Sanchez, C. I., Poza, J., Lopez, M. I., 2009, Hornero, R., Detection of Hard Exudates in Retinal Images Using a Radial Basis Function Classifier., Journals of Biomedical Engineering, No. 7, Vol. 37, 1448-1463.
[11] Listia, R., 2013, Klasifikasi Massa pada Citra Mammogram Berdasarkan Gray Level Cooccurence Matrix (GLCM). Tesis. Jurusan Ilmu Komputer FMIPA UGM, Yogyakarta.
DOI: https://doi.org/10.22146/ijccs.3496
Article Metrics
Abstract views : 6301 | views : 15166Refbacks
- There are currently no refbacks.
Copyright (c) 2014 IJCCS - Indonesian Journal of Computing and Cybernetics Systems
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
View My Stats1