Computer aided diagnosis (CAD) for thorax radiography children patient with segmentation deformable models method

Research output: Contribution to journalConference articlepeer-review

Abstract

This study developed a correlation test Computer Aided Diagnosis (CAD) radiographic of children pulmonary using segmentation Deformable Models method to detect abnormalities. Deformable Models method searched abnormalities by value of the image pixel. Deformable models method used two variations, namely deformable models with median filter and deformable models with wiener filter. Abnormal result lung pixel values of segmentation wiener deformable models with wiener filter is 186-255 and deformable models with median filter is 190-255. Deformable Models used wiener filter have ROC result relatively higher than Deformable models used median filter with value of accuracy 78.5%, sensitivity 74.5%, specificity 80.0%, precision 90.0% and overall error of 21.0%. ROC result of segmentation using Deformable models method evince that Deformable models method can separated normal and abnormal tissue from radiographic of children pulmonary. However, Deformable models method can not definitively determine lung infection from radiographic of children pulmonary but this method can help radiologist to separated normal and abnormal tissue automatically with computer.

Original languageEnglish
Article number012054
JournalIOP Conference Series: Materials Science and Engineering
Volume432
Issue number1
DOIs
Publication statusPublished - 19 Nov 2018
Event1st Materials Research Society-Indonesia Conference and Congress 2017, MRS-INA C and C 2017 - Yogyakarta, Indonesia
Duration: 8 Oct 201712 Oct 2017

Fingerprint Dive into the research topics of 'Computer aided diagnosis (CAD) for thorax radiography children patient with segmentation deformable models method'. Together they form a unique fingerprint.

Cite this