Scripts to combine UNC Shape Analysis with Spherical Wavelet Features
10 likes | 142 Vues
This project outlines the steps to combine UNC shape analysis with spherical wavelet features utilizing tcsh scripts. The process involves two main pipelines: the first includes reading UNC preprocessed META surfaces, re-interpolating the spherical meshes into a recursive icosahedron structure, and applying the itkSWavelet filter to extract spherical wavelet coefficients. The second pipeline focuses on visualizing raw and corrected p-values of these features on mean shape maps. The project aims to enhance shape analysis through wavelet techniques, contributing to insights in neuroimaging.
Scripts to combine UNC Shape Analysis with Spherical Wavelet Features
E N D
Presentation Transcript
Scripts to combine UNC Shape Analysis with Spherical Wavelet Features Plan/Expected Challenges/Publication Team Yi Gao, GT (algorithms) Delphine Nain, GT (algorithms) Martin Styner, UNC (algorithms) • Software: tcsh scripts • PIPELINE 1: (after UNC preprocessing, before UNC statistical scripts) • INPUT: Read UNC preprocessed META surfaces • Re-interpolate the spherical meshes to use a recursive icosahedron structure. The output are remeshed META surfaces. • Run the itkSWavelet filter on re-interpolated meshes to obtain spherical wavelet coefficients • OUTPUT: Write spherical wavelet coefficients (SWC) to a text file that will be read by the UNC statistics scripts • PIPELINE 2: (after UNC stats scripts) • INPUT: raw and corrected p-values for the SWC features • Write visualization scripts to visualize the SWC p-value map on the mean shape • OUTPUT: Colormap that can be visualized KWVisu To be Accomplished by end of Programming Week -Write and test scripts on the female caudate structure -Submit code to NA-MIC repository -Write ITK Insight journal