International Journal on Advanced Science, Engineering and Information Technology, Vol. 6 (2016) No. 2, pages: 239-244, DOI:10.18517/ijaseit.6.2.763

Infiltrate Object Extraction in X-ray Image by Using Math-Morphology Method and Feature Region Analysis

Julius Santony, Jufriadif Na`am


Infiltrate is often called as pulmonary vlek for there are white spotteds on the lung. White spotted could be in form of liquid, condensation, or uncircumcised. The liquid is emerge from blood or suppuration. To detect the existence of infiltrate on the lung, it could be done by doing X-ray Thorax checkup. To observe the infiltrate on x-ray thorax image is unable to be seen by every people, but it is done by experts such as radiologists or the pulmonary experts by doing conscientious research. The research was done by extracting the infiltrate object on x-ray thorax image of tuberculosis patient to clarify the object. The research stage that was done on x-ray thorax image is by detecting the object with segmentation morphology process which consist of dilation and erosion morphology, and side detection by decreasing the value of dilation and erosion morphology. The next stage is extracting the infiltrate object by using binarization and feature region analysis to ommit the unspotted part and determine the infiltrate object from the amount of existing objects. The result of infiltrate object extraction, then, is being calculated for both number and width of each side of lung by using feature region analysis. The result indicates that infiltrate object extraction is able to show an image with the explicit infiltrate object. The result trials of 40 x-ray thorax image on tuberculosis patients proved that the well-extracted images are able to be determined whether on its position, total, and width of infiltrates on lung. The trials of 2 x-ray thorax image on healthy patients are also done as comparisons, and the result indicates that there is no infiltrate objects on both sides of lung.


infiltrate; tuberculosis; x-ray thorax; extracting; morphology; feature region

Viewed: 558 times (since Sept 4, 2017)

cite this paper     download