Filter Selection and Feature Extraction to Distinguish Types of CT Scan Images

Abstract

The study aims to select the most powerful filtering method as input for feature extraction to distinguish the types of Head CT Scan images. Visually determining the scanned medical image (head CT Scan) has difficulty because it has similar results. So that research is needed that aims to determine the types of digital images scanned by using image processing methods, filtering, and feature extraction. This research used a medical image taken from the head CT-Scan of the patient. To be processed using a computer, the data is scanned to obtain digital image data. Furthermore, various filtering methods were selected, such as median, bandpass filter, XYZ colour transformer filter, enhanced local contrast filter, and histogram equalization. The most significant filtered image results are then segmented with the graph cut segmentation method and extracted using the statistical feature extraction method. The results showed that histogram equalization and enhanced local contrast filter methods were the most significant filtering methods. While the mean and standard deviation are the two most important characteristics that can distinguish the three classes of head CT Scan