ANALISIS TEKSTUR CITRA MIKROSKOPIS KANKER PARU MENGGUNAKAN METODE GRAY LEVEL CO-OCCURANCE MATRIX (GLCM) DAN TRANFORMASI WAVELET DENGAN KLASIFIKASI NAIVE BAYES
| Main Authors: | Syifa, Rizky Ayomi; Departemen Fisika, Fakultas Sains dan Matematika, Universitas Diponegoro, Semarang, Adi, Kusworo; Departemen Fisika, Fakultas Sains dan Matematika, Universitas Diponegoro, Semarang, Widodo, Catur Edi; Departemen Fisika, Fakultas Sains dan Matematika, Universitas Diponegoro, Semarang |
|---|---|
| Format: | Article info application/pdf eJournal |
| Bahasa: | eng |
| Terbitan: |
Jurusan Fisika, Fakultas Sains dan Matematika, Universitas Diponegoro
, 2016
|
| Subjects: | |
| Online Access: |
http://ejournal-s1.undip.ac.id/index.php/bfd/article/view/14135 http://ejournal-s1.undip.ac.id/index.php/bfd/article/view/14135/13669 |
Daftar Isi:
- This research, conduct the lung cancer detection system on a microscopic image. The microscopic image used is the result from lung biopsy. If there is a cancerous tissue in the image of lung biopsy, the texture will be irregular, while the image of the normal lung biopsy will have a regular texture. The purpose of this reserach is to develope the lung cancer detection system and also to count the performance of the lung cancer detection system. The clasification process uses two methods, Gray Level Co-Occurance Matrix (GLCM) and Daubechies Wavelet Transform (db1). The Daubechies wavelet transformation is used in decomposition in level 4, while the offset of GLCM is 6. The feature extraction process is done in the transformation wavelet using the 4 subbands, approximation, horizontal Detils coefficients, vertical Detil coefficients and diagonal Detil coefficients, and the the feature extraction of GLCM uses the contrast, correlation, homogenity and energi as the parameter. The naïve bayes classification requires 2 parameter input, do a classification is 4 combination from each method of feature extraction. The result of this research is to extent the level of accuracy for the extraction of the feature extraction in 71,42% wavelet transformation method for the combination coefficients approximation-diagonal Detil coefficients and 80% accuration of GLCM method for the combination of homogeneity-correlation.