Computerized Analysis of Classification of Lung Nodules and Comparison between Homogeneous and Heterogeneous Ensemble of Classifier Model
- 1 December 2011
- conference paper
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE) in 2011 Third National Conference on Computer Vision, Pattern Recognition, Image Processing and Graphics
- p. 231-234
- https://doi.org/10.1109/ncvpripg.2011.56
Abstract
In this paper, we convert multi class subjective ratings for lung nodules from radiologists to binary class problem and use that to classify. We also evaluate the difference in performance between homogenous and heterogeneous ensemble of classifiers. We have used radiologist's characteristic ratings for nodule as subjective feature and extracted 54 low level image features. Instead of predicting nodule characteristic we have used the ground truth provided by radiologists and converted the problem into binary class. Results show that the proposed heterogeneous ensemble classifier model works better than the previous traditional models.Keywords
This publication has 8 references indexed in Scilit:
- Classifier ensembles for fMRI data analysis: an experimentMagnetic Resonance Imaging, 2010
- Random Subspace Ensembles for fMRI ClassificationIEEE Transactions on Medical Imaging, 2010
- Predicting Radiological Panel Opinions Using a Panel of Machine Learning ClassifiersAlgorithms, 2009
- A Two-Step Approach for Feature Selection and Classifier Ensemble Construction in Computer-Aided DiagnosisPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2008
- Combining Pattern ClassifiersPublished by Wiley ,2004
- Is Combining Classifiers with Stacking Better than Selecting the Best One?Machine Learning, 2004
- Combining Classifiers with Meta Decision TreesMachine Learning, 2003
- Computerized Analysis of the Likelihood of Malignancy in Solitary Pulmonary Nodules with Use of Artificial Neural NetworksRadiology, 2000