Min Zhang PhD, Presenter: Nothing to Disclose

Xiangrong Zhou PhD, Abstract Co-Author: Nothing to Disclose

Satoshi Goshima MD, PhD, Abstract Co-Author: Nothing to Disclose

Huayue Cheng, Abstract Co-Author: Nothing to Disclose

Chisako Muramatsu PhD, Abstract Co-Author: Nothing to Disclose

Takeshi Hara MD,PhD, Abstract Co-Author: Nothing to Disclose

Ryujiro Yokoyama, Abstract Co-Author: Nothing to Disclose

Masayuki Kanematsu MD, Abstract Co-Author: Consultant, DAIICHI SANKYO Group

Hiroshi Fujita PhD, Abstract Co-Author: Nothing to Disclose

Computer-aided diagnosis for pulmonary emphysema is very important. The purpose of this study is to propose a reliable and high performance computational method for the classification of different kinds of pulmonary emphysema and normal pulmonary tissues.

A novel pulmonary emphysema classification method is proposed in a statistics way of image morphological features using a texton learning via sparse representation. The proposed method mainly includes four stages: 1. With ROI (Region of Interest) image pre-processing, ROI images are normalized to have zero mean and unit standard deviation; 2. An overcomplete dictionary of textons is constructed via sparse representation; 3. texton feature histograms are learned from the training set; 4. Finally, a ROI image can be classified into the corresponding class by a classifier using the feature histogram.

The proposed scheme was applied to 18 patient cases of non-contrast CT images. Each CT image covers the whole torso region with an isotopic spatial resolution of 0.63 [mm] and a 12 [bits] density resolution. The test images were obtained from 18 different subjects, including 9 healthy subjects and 9 subjects with three subtypes of pulmonary emphysema, including Panlobular emphysema (PLE), Paraseptal emphysema (PSE), centrilobular emphysema (CLE) in different stage. Totally 1984 64×64 region of interests (ROIs) are extracted from the 9 healthy subjects and 1856 64×64 ROIs are extracted from 9 subjects with emphysema. In the experiments, training set was constructed only by 80 ROIs, which account for 4.2% of totally ROIs, for the healthy subjects and subjects with emphysema separately. The classification results are well in accordance with the diagnostic findings performed by doctors in HRCT. the proposed method achieve good ROI classification accuracies around 99% with parameter optimization.

Pulmonary emphysema classification based on texton learning via sparse representation is a promising method for computer-aided diagnosis of the emphysema disease. With this methods, the classification accuracy is steadily improved.

This study shows that texture classification method based on texton learning via sparse representation is very useful for the computer-aided diagnosis in pulmonary emphysema classification in CT image

Zhang, M, Zhou, X, Goshima, S, Cheng, H, Muramatsu, C, Hara, T, Yokoyama, R, Kanematsu, M, Fujita, H, Computer-aided Diagnosis for Pulmonary Emphysema Classification Based on Texton Learning via Sparse Representation.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12031551.html