TY - GEN
T1 - Towards quantification of interstitial pneumonia patterns in lung multidetector CT
AU - Korfiatis, Panayiotis
AU - Karahaliou, Anna
AU - Kazantzi, Alexandra
AU - Kalogeropoulou, Cristina
AU - Costaridou, Lena
PY - 2008
Y1 - 2008
N2 - Quantification of Diffuse Parenchyma Lung Disease (DPLD) patterns challenges Computer Aided Diagnosis schemes in Computed Tomography (CT) lung analysis. In this study, an automated scheme for volumetric quantification of Interstitial Pneumonia (IP) patterns, a subset of DPLDs, is presented, utilizing a MultiDetector CT (MDCT) data set. Initially, Lung Field (LF) segmentation is achieved by 3D automated gray level thresholding combined to wavelet highlighting, followed by a texture based border refinement step. The vessel tree volume is identified and removed from LF, resulting in Lung Parenchyma (LP) volume. Following, the abnormal LP is differentiated from normal LP utilizing a 2 class k-means clustering. Quantification of IP patterns is formulated as a three-class pattern recognition problem to classify abnormal LP into ground glass, reticular and honeycomb patterns, by means of SVM voxel classification, exploiting 3D co-occurrence features. Performance of the proposed scheme in segmenting LF, as well as in quantifying normal LP, ground glass, reticular and honeycomb patterns was evaluated by means of volume overlap on 5 MDCT scans. Volume overlap for left LF and right LF was 0.95±0.03 and 0.96±0.02 respectively, while for normal LP, ground glass, reticular and honeycombing patterns was 0.89±0.02, 0.70±0.04, 0.72±0.05 and 0.71±0.03, respectively.
AB - Quantification of Diffuse Parenchyma Lung Disease (DPLD) patterns challenges Computer Aided Diagnosis schemes in Computed Tomography (CT) lung analysis. In this study, an automated scheme for volumetric quantification of Interstitial Pneumonia (IP) patterns, a subset of DPLDs, is presented, utilizing a MultiDetector CT (MDCT) data set. Initially, Lung Field (LF) segmentation is achieved by 3D automated gray level thresholding combined to wavelet highlighting, followed by a texture based border refinement step. The vessel tree volume is identified and removed from LF, resulting in Lung Parenchyma (LP) volume. Following, the abnormal LP is differentiated from normal LP utilizing a 2 class k-means clustering. Quantification of IP patterns is formulated as a three-class pattern recognition problem to classify abnormal LP into ground glass, reticular and honeycomb patterns, by means of SVM voxel classification, exploiting 3D co-occurrence features. Performance of the proposed scheme in segmenting LF, as well as in quantifying normal LP, ground glass, reticular and honeycomb patterns was evaluated by means of volume overlap on 5 MDCT scans. Volume overlap for left LF and right LF was 0.95±0.03 and 0.96±0.02 respectively, while for normal LP, ground glass, reticular and honeycombing patterns was 0.89±0.02, 0.70±0.04, 0.72±0.05 and 0.71±0.03, respectively.
KW - 3D co-occurrence
KW - Computer aided diagnosis
KW - Diffuse lung diseases quantification
KW - Multidetector CT
KW - Support vector machine classification
UR - http://www.scopus.com/inward/record.url?scp=67549122631&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=67549122631&partnerID=8YFLogxK
U2 - 10.1109/BIBE.2008.4696813
DO - 10.1109/BIBE.2008.4696813
M3 - Conference contribution
AN - SCOPUS:67549122631
SN - 9781424428458
T3 - 8th IEEE International Conference on BioInformatics and BioEngineering, BIBE 2008
BT - 8th IEEE International Conference on BioInformatics and BioEngineering, BIBE 2008
T2 - 8th IEEE International Conference on BioInformatics and BioEngineering, BIBE 2008
Y2 - 8 October 2008 through 10 October 2008
ER -