Learning COPD Sensitive Filters in Pulmonary CT

Abstract

Abstract. The standard approaches to analyzing emphysema in computed tomography (CT) images are visual inspection and the relative area of voxels below a threshold (RA). The former approach is subjective and impractical in a large data set and the latter relies on a single threshold and independent voxel information, ignoring any spatial correlation in intensities. In recent years, supervised learning on texture features has been investigated as an alternative to these approaches, showing good results. However, supervised learning requires labeled samples, and these samples are often obtained via subjective and time consuming visual scoring done by human experts. In this work, we investigate the possibility of applying supervised learning using texture measures on random CT samples where the labels are based on external, non-CT measures. We are not targeting emphysema directly, instead we focus on learning textural di erences that discriminate subjects with chronic obstructive pulmonary disease (COPD) from healthy smokers, and it is expected that emphysema plays a major part in this. The proposed texture based approach achieves an 69% classication accuracy which is signi cantly better than RA's 55% accuracy.

Image Dissimilarity-Based Quantification of Lung Disease from CT

“…This is compared to using d(·, ·) directly as distance in a k nearest neighbor classifier (kNN), which for k = 1 corresponds to template matching, and to fusing individual ROI classifications, classified using kNN, for image classification [6]. A posterior probability of each image being positive is obtained using leave-one-out estimation, and receiver operating characteristic (ROC) analysis is used to evaluate the different methods by means of the area under the ROC curve (AUC).…”

“…Since the application in this paper is quantification of COPD in pulmonary CT images based on textural appearance in the ROIs, we will focus on image dissimilarity measures suitable for this purpose. In texture-based classification of lung tissue, the texture is sometimes assumed stationary [3,4,6,7]. We will make the same assumption and, therefore, disregard the spatial location of the ROIs within the lungs.…”

“…Quantification of abnormality in medical images often involves classification of regions of interest (ROIs), and combination of individual ROI classification outputs into one global measure of disease for the entire image [1,2,3,4,5,6,7]. These measures may, e.g., express a probability of the presence of certain abnormalities or reflect the extent or severity of disease.…”

“…Moreover, in some cases only global image labels are available, while the images are still represented by ROIs in order to capture localized abnormalities. In some studies, this is handled by propagating the image label to the ROIs within that image, which again allows fusion of individual ROI classifications, to obtain a global image measure [4,5,6]. However, an image showing abnormality will generally comprise both healthy and abnormal regions, and the above approach, incorrectly, labels ROIs without abnormality in such an image as abnormal.…”

“…Several general purpose classifiers built in the obtained image dissimilarity spaces are evaluated and compared to image classification by fusion of individual ROI classifications as was used in [6].…”