Deformation Estimation of Deflated Lung using Kernel Method based on the Relative Position of Some Landmarks

Abstract

It is expected to grasp how the lung deform by the deaeration during surgery compared to the inflated lung. In this study we propose a method to estimate deformation of the deflated lung from the inflated one based on the relative position of some landmarks using dog lungs. The kernel method was employed for the estimation as a machine learning technique. We achieved mean local positional error of 2.96 mm for test data where the volume reduction by the deaeration was 40 %.