Lung Area Segmentation on Chest X-Ray Based on Texture Image Feature

Abstract

Advanced technology has permitted many innovations in computer aided diagnostics. One of the most popular studies done to the medical image is to segment specific body areas such as the lung for further image analysis. The segmentation of lung area in chest X-ray (CXR) based regular techniques such as contour and level set based are popular however these methods are timely and require special initialization process or else it will accidently cause false positive area selection. Therefore, the requirement for a better segmentation method to segment the lung area in CXR image should be highlighted. A quick solution to cope with this obstacle is to propose a noble feature extraction technique based on texture feature using the Gray Level Co-Occurrence Matrix (GLCM) so that image features could be grouped together based on similar feature vectors. Therefore, in this paper we are sharing our experience conducting a lung segmentation experiment using the CXR image. In order to execute the experiment we also shared six processes that are the common method in the lung segmentation task. The segmentation output derived from the experiment shows a promising appearance although it is not accurately similar with the original lung area in the actual image.