Tumors are an abnormally growing cells and are divided into two types, benign and malignant. Cervical cancer is one of the types of malignant tumors. It is a type of cancer with the highest prevalence in the world. Nevertheless, most cervical cancer patients came to doctors with end-stage cancer conditions. Various diagnostic imaging modalities are used to determine the location, the size and the severity of cervical cancer that affects patients, one of them is CT-Scan. There are not many researches on digital image processing of cervical cancer used CT-Scan images so that makes this research different from others. The process of segmentation, feature extraction, and classification are some examples of digital image processing techniques applied to cervical cancer for further analysis. This study created a program which used CT-Scan image data and Region-based Snake Model as the segmentation method. It also used Support Vector Machine (SVM) and feature extraction in the form of Gray Level Co-occurrence Matrix (GLCM) of texture analysis. The segmentation process was used to obtain the region of cervical abnormalities. The classification process was expected to categorize the image data onto normal images and cervical abnormally images. Results of segmentation process which gave limits of the region of the cervical abnormalities was quite appropriate and fit to diagnosis of specialists or radiologists. The classification process gave good statistical results in distinguishing normal images and cervical abnormality images, i.e., sensitivity, specificity, accuracy, precision, overall error, and Area Under Curve (AUC) reached 88%, 92%, 90%, 91.67%, 10%, and 90%.
|Journal of Physics: Conference Series
|Published - 12 Oct 2018
|5th International Conference on Research, Implementation, and Education of Mathematics and Science, ICRIEMS 2018 - Yogyakarta, Indonesia
Duration: 7 May 2018 → 8 May 2018