Thonnes, E., Bhalerao, Abhir and Parr, D. (2010) Classification of lung disease in HRCT scans using integral geometry measures and functional data analysis. In: Proceedings of the Medical Image Understanding and Analysis (MIUA'10), Coventry, UK, 6-7 July, 2010

Abstract

A framework for classification of chronic lung disease from high-resolution CT scans is presented. We use a set of features which measure the local morphology and topology of the 3D voxels within the lung parenchyma and apply functional data classification to the extracted features. We introduce the measures, Minkowski functionals, which derive from integral geometry and show results of classification on lungs containing various stages of chronic lung disease: emphysema, fibrosis and honey-combing. Once trained, the presented method is shown to be efficient and specific at characterising the distribution of disease in HRCT slices

Request changes to a record

Actions (login required)

View Item