Lung Tumor Volumetric Estimation and Segmentation using Adaptive Multiple Resolution Contour Model

Abstract

The need to identify pathology is strongly felt in image guided analysis. Tumor localization is essential for targeting the tumor. Tumor boundaries are often hard to delineate due to high variation in shape and size of the tumor. The tumor boundary is usually defined using the thresholding technique which may have high FP rate due to partial volume effects. Hence, this AMRC technique has to be repeated several times in order to get a more accurate result. The multiple thresholding technique often leads to an undesirable situation where the segmentation boundary contains holes which are hard to fill. The AMRC is a dynamic contour model that is able to adaptively adjust the contour to better fit the contour of the tumor. Segmentation of SVT involves two steps viz. identification of peak intensity pixels with the HECS model and definition of an adaptive contour with AMRC.