wilmsm / localizedssm

A Kernelized Multi-level Localization Method for Flexible Shape Modeling with Few Training Data (MICCAI 2020 paper)

This repository contains Python code for the MICCAI 2020 paper

A Kernelized Multi-level Localization Method for Flexible Shape Modeling with Few Training Data by Matthias Wilms, Jan Ehrhardt, Nils D. Forkert (see https://doi.org/10.1007/978-3-030-59719-1_74)

This code is a Python-based extension of the MATLAB codebase we created for our Medical Image Analysis paper from 2017 (see https://imi.uni-luebeck.de/multi-resolution-multi-object-statistical-shape-models)

See demo_miccai.py as a starting point. If you want to replicate our results (small-scale; see paper), you first need to obtain the SCR/JSRT data. See the MATLAB code from our MedIA paper on how to obtain and convert that data set.

Please cite our papers, if you are using our code for your research:

@inproceedings{miccai2020,
    Author = {Matthias Wilms and Jan Ehrhardt and Nils Daniel Forkert},
    Title = {A Kernelized Multi-level Localization Method for Flexible Shape Modeling with Few Training Data},
    Booktitle = {Medical Image Computing and Computer Assisted Intervention -- {MICCAI 2020}},
    Year = {2020}
}

@article{media2017,
    Title = {Multi-resolution multi-object statistical shape models based on the locality assumption},
    Author = {Matthias Wilms and Heinz Handels and Jan Ehrhardt},
    Journal = {Medical Image Analysis},
    Year = {2017},
    Number = {5},
    Pages = {17--29},
    Volume = {38}
}