Abstract
Radiologists are adept at recognizing the character and extent of lung parenchymal abnormalities in computed tomography (CT) scans. However, the inconsistent differential diagnosis due to subjective aggregation necessitates the exploration of automated classification based on supervised or unsupervised learning. The robustness of supervised learning depends on the training samples. Towards optimizing emphysema classification, we introduce a physician-in-the-loop feedback approach to minimize ambiguity in the selected training samples. An experienced thoracic radiologist selected 412 regions of interest (ROIs) across 15 datasets to represent 124, 129, 139 and 20 training samples of mild, moderate, severe emphysema and normal appearance, respectively. Using multi-view (multiple metrics to capture complementary features) inductive learning, an ensemble of seven un-optimized support vector models (SVM) each based on a specific metric was constructed in less than 6 s. The training samples were classified using seven SVM models and consensus labels were created using majority voting. In the active relearning phase, the ensemble-expert label conflicts were resolved by the expert. The efficacy and generality of active relearning feedback was assessed in the optimized parameter space of six general purpose classifiers across the seven dissimilarity metrics. The proposed just-in-time active relearning feedback with un-optimized SVMs yielded 15 % increase in classification accuracy and 25 % reduction in the number of support vectors. The average improvement in accuracy of six classifiers in their optimized parameter space was 21 %. The proposed cooperative feedback method enhances the quality of training samples used to construct automated classification of emphysematous CT scans. Such an approach could lead to substantial improvement in quantification of emphysema.
Similar content being viewed by others
References
Hurd S: The impact of COPD on lung health worldwide: epidemiology and incidence. Chest 117(2 Suppl):1S–4S, 2000
Hochhegger B: CT of pulmonary emphysema: current status, challenges, and future directions. Eur Radiol 19(7):1696, 2009. author reply 1967
Jemal A, Ward E, Hao Y, Thun M: Trends in the leading causes of death in the United States, 1970–2002. JAMA 294(10):1255–1259, 2005
Larson DB, Johnson LW, Schnell BM, Salisbury SR, Forman HP: National trends in CT use in the emergency department: 1995–2007. Radiology 258(1):164–173, 2011
Muller N, Staples C, Miller R, Abboud R: Density mask. An objective method to quantitate emphysema using computed tomography. CHEST J 94(4):782–787, 1988
Wang Z, Gu S, Leader JK, et al: Optimal threshold in CT quantification of emphysema. Eur Radiol 23(4):975–984, 2013
Marsh S, Aldington S, Williams MV, et al: Utility of lung density measurements in the diagnosis of emphysema. Respir Med 101(7):1512–1520, 2007
Boedeker KL, McNitt-Gray MF, Rogers SR, et al: Emphysema: effect of reconstruction algorithm on CT imaging measures. Radiology 232(1):295–301, 2004
Gierada DS, Bierhals AJ, Choong CK, et al: Effects of CT section thickness and reconstruction kernel on emphysema quantification: relationship to the magnitude of the CT emphysema index. Acad Radiol 17(2):146, 2010
Mendonca PR, Padfield DR, Ross JC, Miller JV, Dutta S, Gautham SM: Quantification of emphysema severity by histogram analysis of CT scans. Med Image Comput Comput Assist Interv 8(Pt 1):738–744, 2005
Prasad M, Sowmya A: Multi-level classification of emphysema in HRCT lung images using delegated classifiers. Med Image Comput Comput Assist Interv 11(Pt 1):59–66, 2008
Xu Y, Sonka M, McLennan G, Guo J, Hoffman EA: MDCT-based 3-D texture classification of emphysema and early smoking related lung pathologies. IEEE Trans Med Imaging 25(4):464–475, 2006
Castaldi PJ, San José Estépar R, Mendoza CS, et al: Distinct quantitative computed tomography emphysema patterns are associated with physiology and function in smokers. Am J Respir Crit Care Med 188(9):1083–1090, 2013
Estepar RSJ, Mendoza C, Ross J, Diaz A, Lynch D, Crapo J: Quantifying patterns of emphysema by local density histogram in chest CT scans. Am J Respir Crit Care Med. 185:A4331, 2012
Nava R, Marcos JV, Escalante-Ramírez B, Cristóbal G, Perrinet LU, Estépar RSJ: Advances in texture analysis for emphysema classification. Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications. Lect Notes Comput Sci, Springer 8259:214–221, 2013
Barandela R, Gasca E. Decontamination of training samples for supervised pattern recognition methods. Advances in Pattern Recognition, Springer; 2000:621–630.
Tuia D, Ratle F, Pacifici F, Kanevski MF, Emery WJ: Active learning methods for remote sensing image classification. Geosci Remote Sens IEEE Trans 47(7):2218–2232, 2009
Tong S, Koller D: Support vector machine active learning with applications to text classification. J Mach Learn Res 2:45–66, 2002
McCallum A, Nigam K. Employing EM in pool-based active learning for text classification. In Proceedings of ICML-98, 15th International Conference on Machine Learning, 1998.
Cortes C, Vapnik V: Support-vector networks. Mach Learn 20(3):273–297, 1995
Bishop CM: Pattern Recognition and Machine Learning, vol. 1. Springer, New York, 2006
Burges CJ: A tutorial on support vector machines for pattern recognition. Data Min Knowl Discov 2(2):121–167, 1998
Shaker S, Dirksen A, Laursen L, et al: Short‐term reproducibility of computed tomography‐based lung density measurements in alpha‐1 antitrypsin deficiency and smokers with emphysema. Acta Radiol 45(4):424–430, 2004
Madani A, Zanen J, De Maertelaer V, Gevenois PA: Pulmonary emphysema: objective quantification at multi-detector row CT-comparison with macroscopic and microscopic morphometry. Radiology 238(3):1036–1043, 2006
Suykens JAK, Vandewalle J: Least squares support vector machine classifiers. Neural Process Lett 9(3):293–300, 1999
Mierswa I, Wurst M, Klinkenberg R, Scholz M, Euler T. Yale: Rapid prototyping for complex data mining tasks. In Proceedings of the 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2006.
Polyak R, Ho SS, Griva I: Support vector machine via nonlinear rescaling method. Optim Lett 1(4):367–378, 2007
de Torres JP, Bastarrika G, Zagaceta J, et al: Emphysema presence, severity, and distribution has little impact on the clinical presentation of a cohort of patients with mild to moderate COPD. Chest 139(1):36–42, 2011
Muslea I, Minton S, Knoblock CA: Active learning with multiple views. J Artif Intell Res 27:203–233, 2006
Heisele B, Serre T, Prentice S, Poggio T: Hierarchical classification and feature reduction for fast face detection with support vector machines. Pattern Recogn 36(9):2007–2017, 2003
Habib T, Inglada J, Mercier G, Chanussot J: Support vector reduction in SVM algorithm for abrupt change detection in remote sensing. IEEE Geosci Remote Sens 6(3):606–610, 2009
Sørensen L, Shaker SB, De Bruijne M. Texture classification in lung CT using local binary patterns. Medical Image Computing and Computer-Assisted Intervention–MICCAI 2008: Springer; 2008:934–941.
Li Y, Maguire L: Selecting critical patterns based on local geometrical and statistical information. IEEE Trans Pattern Anal Mach Intell 33(6):1189–1201, 2011
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Raghunath, S., Rajagopalan, S., Karwoski, R.A. et al. Active Relearning for Robust Supervised Training of Emphysema Patterns. J Digit Imaging 27, 548–555 (2014). https://doi.org/10.1007/s10278-014-9686-z
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10278-014-9686-z