Pathological lung classification using random forest classifier

The correct classification of pathological lung images could be a little bit complex task in medical imaging applications. The present day analyses aren't correct to afford an improved solution for images with dense pathologies. In this proposed framework the patch approximation has been performed with the lung Computed Tomography (CT) scan images. Next, the patches are labeling with corresponding categories. The feature vectors like Gray Level Co-occurrence Matrix (GLCM), Gray Level Run Length Matrix (GLRLM) and Rotation-invariant Gabor Local Binary Pattern (RGLBP) are extracted from the patches. Then, these feature vectors are given to the random forest classifier. The various classes consider for the analysis are Normal, Ground-glass opacity, Honeycomb and Tree-in-bud. Finally the performance of the proposed technique is evaluated by means of sensitivity, specificity and accuracy.