ANALISA FITUR TEKSTUR NUKLEUS DAN DETEKSI SITOPLASMA PADA CITRA PAP SMEAR

Abstract

Currently, the identification of Pap smear cells in the early detection process of cervical cancer is still an important stage of the process. The ease of detecting Pap smear cells will be very helpful in the introduction of cell abnormalities. Pap smear cell images consist of parts of the nucleus and cytoplasm. Proper analysis of parts of the nucleus and cytoplasm will facilitate the process of identifying cell abnormalities. This study presents Pap smear cell texture analysis on the pap smear cell nucleus and segmentation of the cytoplasmic area. Texture analysis was performed on 250 cell images of the nucleus. While cytoplasmic segmentation was performed for 887 cytoplasmic cell images. Senua cell image used has class categories categorized into seven classes. Three classes of them are normal cell image class categories that include: Normal Superficial, Normal Intermediate, and Normal Columnar, and the other four classes are abnormal cell image class categories which include: mild dysplasia, moderate dysplasia, severe dysplasia, and carcinoma Di There. The method used for texture analysis using 8-bit grayscale. And using the second sequence of Gray Level Co-occurrence Matrix (GLCM) statistics, with contrast, correlation, energy, homogeneity, and entropy features. Cytoplasmic detection uses edge detection and some morphological analyzes. The results showed that the numerical results of all the texture of the nucleus for each class of Pap smear image had slightly different properties. As for the results of cytoplasmic detection showed that the stage of the proposed detection process results in a clean area of the cytoplasm and can be detected well