An image of Josua Sassen

Hello there! I am an applied mathematician currently working as a MathInGreaterParis Postdoctoral Fellow at the Centre Borelli of the ENS Paris-Saclay, where I am mentored by Alain Trouvé. Previously, I obtained my BSc, MSc and PhD in Mathematics from the University of Bonn, the latter two advised by Martin Rumpf.

My research is concerned with variational problems in geometry processing. Broadly speaking, I am interested in numerical methods for such problems arising from shape spaces and shape optimization. These typically find applications in computer graphics and geometric design. Recently, I have also become interested in using techniques from machine learning to tackle these problems. If you would like to chat, feel free to send me an email!

News

Publications
Peer-Reviewed
  1. An Elastic Basis for Spectral Shape Correspondence
    Florine Hartwig, Josua Sassen, Omri Azencot, Martin Rumpf, Mirela Ben-Chen. ACM SIGGRAPH 2023 Conference Proceedings. 2023.
    @inproceedings{HaSaAz23,
    author = {Hartwig, Florine and Sassen, Josua and Azencot, Omri and Rumpf, Martin and Ben-Chen, Mirela},
    title = {{An Elastic Basis for Spectral Shape Correspondence}},
    year = {2023},
    isbn = {9798400701597},
    publisher = {Association for Computing Machinery},
    address = {New York, NY, USA},
    doi = {10.1145/3588432.3591518},
    booktitle = {ACM SIGGRAPH 2023 Conference Proceedings},
    articleno = {58},
    numpages = {11},
    location = {Los Angeles, CA, USA},
    series = {SIGGRAPH '23}
    }
    PDF doi Supplement Code Teaser Talk by Florine
  2. Parametrizing Product Shape Manifolds by Composite Networks
    Josua Sassen, Klaus Hildebrandt, Martin Rumpf, and Benedikt Wirth. International Conference on Learning Representations. 2023. spotlight paper (notable top 25%).
    @article{SaHiRu23,
    title={{Parametrizing Product Shape Manifolds by Composite Networks}},
    author={Sassen, Josua and Hildebrandt, Klaus and Rumpf, Martin and Wirth, Benedikt},
    journal={International Conference on Learning Representations},
    year={2023},
    archiveprefix = {arXiv},
    eprint = {2302.14665},
    url={https://openreview.net/forum?id=F_EhNDSamN}
    }
    PDF arXiv OpenReview Freaky Torus
  3. A Pessimistic Bilevel Stochastic Problem for Elastic Shape Optimization
    Johanna Burtscheidt, Matthias Claus, Sergio Conti, Martin Rumpf, Josua Sassen, Rüdiger Schultz. Mathematical Programming. 2023.
    @article{BuClCo23,
    author = {Burtscheidt, Johanna and Claus, Matthias and Conti, Sergio and Rumpf, Martin and Sassen, Josua and Schultz, R{\"u}diger},
    title = {A {P}essimistic {B}ilevel {S}tochastic {P}roblem for {E}lastic {S}hape {O}ptimization},
    journal={Mathematical Programming},
    year={2023},
    volume={198},
    number={2},
    pages={1125--1151},
    issn={1436-4646},
    doi={10.1007/s10107-021-01736-w}
    }
    PDF doi arXiv Code
  4. Association of Reading Performance in Geographic Atrophy Secondary to Age-Related Macular Degeneration With Visual Function and Structural Biomarkers
    Sandrine H. Künzel, Moritz Lindner, Josua Sassen, Philipp T. Möller, Lukas Goerdt, Matthias Schmid, Steffen Schmitz-Valckenberg, Frank G. Holz, Monika Fleckenstein, Maximilian Pfau. JAMA Ophthalmology. 2021.
    @article{KuLiSa21,
    author = {Künzel, Sandrine H. and Lindner, Moritz and Sassen, Josua and Möller, Philipp T. and Goerdt, Lukas and Schmid, Matthias and Schmitz-Valckenberg, Steffen and Holz, Frank G. and Fleckenstein, Monika and Pfau, Maximilian},
    title = {{Association of Reading Performance in Geographic Atrophy Secondary to Age-Related Macular Degeneration With Visual Function and Structural Biomarkers}},
    journal = {JAMA Ophthalmology},
    volume = {139},
    number = {11},
    pages = {1191-1199},
    year = {2021},
    issn = {2168-6165},
    doi = {10.1001/jamaophthalmol.2021.3826}
    }
    doi
  5. A Phase-field Approach to Variational Hierarchical Surface Segmentation
    Janos Meny, Martin Rumpf, Josua Sassen. Computer Aided Geometric Design. 2021.
    @article{MeRuSa21,
    title = {{A Phase-field Approach to Variational Hierarchical Surface Segmentation}},
    journal = {Computer Aided Geometric Design},
    volume = {89},
    pages = {102025},
    year = {2021},
    issn = {0167-8396},
    doi = {10.1016/j.cagd.2021.102025},
    author = {Janos Meny and Martin Rumpf and Josua Sassen}
    }
    PDF doi Code
  6. Nonlinear Deformation Synthesis via Sparse Principal Geodesic Analysis
    Josua Sassen, Klaus Hildebrandt, Martin Rumpf. Computer Graphics Forum (Proc. SGP). 2020.
    @article{SaHiRu20,
    author = {Sassen, Josua and Hildebrandt, Klaus and Rumpf, Martin},
    title = {{Nonlinear Deformation Synthesis via Sparse Principal Geodesic Analysis}},
    journal = {Computer Graphics Forum (Proc. SGP)},
    year = {2020},
    volume = {39},
    number = {5},
    pages = {119-132},
    doi = {10.1111/cgf.14073}
    }
    PDF doi Talk Video
  7. Geometric optimization using nonlinear rotation-invariant coordinates
    Josua Sassen, Behrend Heeren, Klaus Hildebrandt, Martin Rumpf. Computer Aided Geometric Design. 2020.
    @article{SaHeHi20,
    author = {{Sassen}, Josua and {Heeren}, Behrend and {Hildebrandt}, Klaus and {Rumpf}, Martin},
    title = {{Geometric optimization using nonlinear rotation-invariant coordinates}},
    journal = {Computer Aided Geometric Design},
    volume = {77},
    pages = {101829},
    year = {2020},
    issn = {0167-8396},
    doi = {10.1016/j.cagd.2020.101829}
    }
    PDF doi arXiv
Other
  1. Riemannian Calculus and Shape Optimization on the Space of Discrete Surfaces
    Josua Sassen. Dissertation, University of Bonn. 2023.
    @phdthesis{Sa23,
    author = {{Josua Raphael Sassen}},
    title = {{Riemannian Calculus and Shape Optimization on the Space of Discrete Surfaces}},
    school = {Rheinische Friedrich-Wilhelms-Universität Bonn},
    year = 2023,
    url = {https://hdl.handle.net/20.500.11811/10960}
    }
    PDF doi
  2. Solving Variational Problems Using Nonlinear Rotation-invariant Coordinates
    Josua Sassen, Behrend Heeren, Klaus Hildebrandt, Martin Rumpf. Symposium on Geometry Processing 2019 – Posters. 2019.
    @inproceedings {SaHeHi19,
    booktitle = {Symposium on Geometry Processing 2019 -- Posters},
    title = {{Solving Variational Problems Using Nonlinear Rotation-invariant Coordinates}},
    author = {Sassen, Josua and Heeren, Behrend and Hildebrandt, Klaus and Rumpf, Martin},
    year = {2019},
    publisher = {The Eurographics Association},
    DOI = {10.2312/sgp.20191213}
    }
    doi
  3. Discrete Gauß–Codazzi Equations for Efficient Triangle Mesh Processing
    Josua Sassen. Master's Thesis, University of Bonn. 2019.
    @MastersThesis{Sa19,
    author = {Sassen, Josua},
    title = {Discrete {Gau{\ss}--Codazzi} Equations for Efficient Triangle Mesh Processing},
    year = {2019},
    type = {Master's Thesis},
    school = {University of Bonn}
    }
    PDF
Code & Data
The logo of GOAST: A triangulated Pac-Man ghost.
Geometric Optimization And Simulation Toolbox
GOAST

The Geometric Optimization and Simulation Toolbox (GOAST) is a C++ library for research in Geometry Processing. It primarily provides tools for and examples of variational approaches on triangle meshes. Notably, it includes numerical methods for the Riemannian shape space of discrete shells.

@software{GOAST,
title = {{The Geometric Optimization And Simulation Toolbox (GOAST)}},
author = {Heeren, Behrend and Sassen, Josua},
url = {https://gitlab.com/numod/goast},
year = {2020}
}
GitLab
Elastic energies for shells have two components: membrane and bending distortions.
Nonlinear Discrete Shell Energies
The Easy Way

This C++ package provides a stand-alone way to use the nonlinear shell deformation energies from GOAST with triangle meshes described by matrices as used by libigl. Furthermore, it provides convenient interfaces for MATLAB (gptoolbox) and Python (libigl).

@software{HeSa23,
author = {Heeren, Behrend and Sassen, Josua},
publisher = {bonndata},
title = {{An Implementation of Nonlinear Discrete Shell Energies}},
year = {2023},
version = {V1},
doi = {10.60507/FK2/YMRPMF},
url = {https://gitlab.com/numod/shell-energy}
}
GitLab bonndata
A visualization of the factors of the Freaky Torus shape space.
Freaky Torus

This is the Python code to compute samples on our shape space Freaky Torus of deformed tori, a synthetic shape space with factors S1×S1×T2, as used in our ICLR 2023 paper. Additional to the code, it also includes the exact dataset that we used for the results of our paper.

@data{SaHiRu23b,
author = {Sassen, Josua and Hildebrandt, Klaus and Rumpf, Martin and Wirth, Benedikt},
publisher = {bonndata},
title = {{The Freaky Torus}},
year = {2023},
version = {V1},
doi = {10.60507/FK2/LORXU7},
url = {https://gitlab.com/jrsassen/freaky-torus}
}
GitLab bonndata Paper