Eigenfunction Expansion Surface Mesh Smoothing of 3D Pleural Plaques' Model from CT Data

For non-invasive diagnostics of pleural mesothelioma in its early stage, physicians inspect CT-data to find pleural plaques which is time-consuming and underlies inter- and intra-reader variability. To increase the accuracy of the localization and the topological finding of these small areas, 3D CT images are processed through a computer-aided diagnosis system that delineates plaques. In order to follow the development of each plaque, thin plate spline interpolation was used to calculate the volume, but cannot handle a complex branch structure. The surface smoothing algorithm based on Laplace-Beltrami eigenfunction expansion has been experimented to create an even closer-to-reality 3D model of plaques.