Journal article // Jurnal TICOM






Ekstraksi Dan Klasifikasi Tekstur Citra Sel Nukleus Pap Smear
2013
Dwiza Riana, Dwi Widyantoro, Tati Latifah R. Mengko

Metrik

  • Eye Icon 301 kali dilihat
  • Download Icon 495 downloads
Metrics Icon 301 kali dilihat  //  495 downloads
Abstrak

— This paper presents a texture analysis and comparison of clasification of cell nucleus images. Texture analysis will be focused on the nuclei of Image Pap smear cell. The method of analysis texture is the statistical second order of Grey Level Co-occurrence Matrix (GLCM). There are five parameter that will be extracted, viz. contrast, correlation, energy, homogeneity and entropy. The image nuclei used in this work are cropped images from Herlev data bank. The images from 917 subjects are categorized into seven classes. Three classes of which are normal cell image class categories that include: Normal Superficial, Normal Intermediate, and Normal Columnar, and the other four classes are categories of abnormal cell image class that include: Mild (Light) Dysplasia, Moderate Dysplasia, Severe Dysplasia and Carcinoma In Situ. The process of texture analysis developed using grayscale 8 bit level. The preprocessing of images will be conducted before texture analysis in order to improve contrast in nuclei. Based on the numerical result of all parameter, class normal and abnormal of Pap smear image have slightly different properties for correlation, energy, homogeneity and entropy. Originally, there have been 18 fetures of texture which were created to classify into two classes by decision tree classifier, ie normal and abnormal cell. The experimental study shows that in two-class classification, normal and abnormal based on the texture features and using the Decision Tree learning algorithm (J48) classifiers with the Weka Correctly Classification Instances (CCI) and Kappa Coefficient classification performance measures, the Decision Tree learning algorithm (J48) classifier performs the best with the CCI of 73.8277% and the Kappa Coefficient of 0.2785. Keywords — Pap smear images, nucleus, texture, GLCM, classification, decision tree, cervical cancer

Full text
Show more arrow
 

Metrik

  • Eye Icon 301 kali dilihat
  • Download Icon 495 downloads
Metrics Icon 301 kali dilihat  //  495 downloads