TY - GEN
T1 - Identification of Pulmonary Lung Nodules Ce ntroid on CT Scans Using Moment Analysis
AU - Tjahyaningtijas, Hapsari Peni Agustin
AU - Purnama, I. Ketut Eddy
AU - Purnomo, Mauridhi H.
N1 - Publisher Copyright:
© 2018 IEEE.
PY - 2018/10/16
Y1 - 2018/10/16
N2 - Lung cancer becomes one of the diseases with a high mortality rate. The identification of lung nodules as a representation of lung cancer becomes a very important part. One method of identifying lung nodules is to segment the lung areas. Increased performance of segmentation is a thing that developed by experts. This research is a preliminary study that develops one of the lung nodule identification techniques by determining its centroid. This centroid will give information how accurate the segmentation method is. The best segmentation method so far gives the smallest value shift of centroid to the ground truth. In determining the central point of the lung nodule, lung cancer CT image is segmented using three different methods namely edge active contour, active base region contour and growing region. The segmented area is then analyzed using a moment analysis that expressed by the center of mass. Center of mass represents centroid coordinates. The coordinate shift between ground truth and suspected nodule coordinates expressed by the standard error value and euclidian distance in pixels. The smallest shift gives a smallest standard error and Euclidean distance. The results showed that the region growing method resulted in 0.004327182 for standard error value and 3.821609503 for a euclidian distance which is smallest than others segmentation methods.
AB - Lung cancer becomes one of the diseases with a high mortality rate. The identification of lung nodules as a representation of lung cancer becomes a very important part. One method of identifying lung nodules is to segment the lung areas. Increased performance of segmentation is a thing that developed by experts. This research is a preliminary study that develops one of the lung nodule identification techniques by determining its centroid. This centroid will give information how accurate the segmentation method is. The best segmentation method so far gives the smallest value shift of centroid to the ground truth. In determining the central point of the lung nodule, lung cancer CT image is segmented using three different methods namely edge active contour, active base region contour and growing region. The segmented area is then analyzed using a moment analysis that expressed by the center of mass. Center of mass represents centroid coordinates. The coordinate shift between ground truth and suspected nodule coordinates expressed by the standard error value and euclidian distance in pixels. The smallest shift gives a smallest standard error and Euclidean distance. The results showed that the region growing method resulted in 0.004327182 for standard error value and 3.821609503 for a euclidian distance which is smallest than others segmentation methods.
KW - Centroid
KW - Edge Active Contour
KW - Moment Analysis
KW - Region Base Active Contour
KW - region Growing
UR - http://www.scopus.com/inward/record.url?scp=85056811814&partnerID=8YFLogxK
U2 - 10.1109/ICoIAS.2018.8493754
DO - 10.1109/ICoIAS.2018.8493754
M3 - Conference contribution
AN - SCOPUS:85056811814
T3 - 2018 International Conference on Intelligent Autonomous Systems, ICoIAS 2018
SP - 67
EP - 71
BT - 2018 International Conference on Intelligent Autonomous Systems, ICoIAS 2018
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2018 International Conference on Intelligent Autonomous Systems, ICoIAS 2018
Y2 - 1 March 2018 through 3 March 2018
ER -