Timo Milbich Ph.D.

I actively contribute to research on Computer Vision and (Deep) Machine Learning as a Postdoctoral Researcher in the Machine Vision and Learning Group at LMU Munich. My interest in and passion for processing and analyzing visual data sparked during my studies in Mathematics and Scientifc Computing, which I fully dedicated to the areas of Computer Vision, Machine Learning and Optimization. Since then I have been able to explore these areas further while pursuing a Ph.D. in Computer Vision at the Heidelberg Collaboratory for Image Processing (HCI) - one of Germany’s biggest and prestigious institutes in this area. My work focuses on learning representations for and similarities between images, as well as analyzing human poses and their dynamics.

Aside from my research and teaching duties, I love to play Volleyball with my team and friends, hike and get lost in nature around the globe or just spend my time in cafes reading books and learning new things.

news

Mar 10, 2022	❗ We will give a tutorial on Deep Visual Similarity and Metric Learning at CVPR’22. ❗
Sep 28, 2021	Paper accepted at NeuRIPS 2021.
Jul 22, 2021	Paper accepted at ICCV 2021.
Feb 28, 2021	Three papers accepted at CVPR 2021.

selected publications

(For full list, please see 'publications' or Google Scholar profile.)

DiVA: Diverse Visual Feature Aggregation for Deep Metric Learning

Milbich*, T., Roth*, K., Bharadhwaj, H., Sinha, S., Bengio, Y., Ommer, B., and Cohen, J. P.

In European Conference on Computer Vision (ECCV) 2020

ECCV

arXiv Bib Code

@inproceedings{milbich2020diva,
  abbr = {ECCV},
  title = {DiVA: Diverse Visual Feature Aggregation for Deep Metric Learning},
  author = {Milbich*, T. and Roth*, K. and Bharadhwaj, H. and Sinha, S. and Bengio, Y. and Ommer, B. and Cohen, J. P.},
  year = {2020},
  booktitle = {European Conference on Computer Vision (ECCV)},
  selected = {true},
  bibtex_show = {true},
  arxiv = {2004.13458},
  thumbnail = {/assets/img/eccv21_thumbnail.png},
  code = {https://github.com/Confusezius/ECCV2020_DiVA_MultiFeature_DML}
}

Sharing Matters for Generalization in Deep Metric Learning

Milbich*, T., Roth*, K., Brattoli, B., and Ommer, B.

IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) 2020

TPAMI

arXiv Bib HTML

@article{milbich2020sharing,
  abbr = {TPAMI},
  title = {Sharing Matters for Generalization in Deep Metric Learning},
  author = {Milbich*, T. and Roth*, K. and Brattoli, B. and Ommer, B.},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)},
  year = {2020},
  publisher = {IEEE},
  selected = {true},
  arxiv = {2004.05582},
  html = {https://www.computer.org/csdl/journal/tp/5555/01/09141449/1lu2CexRydG},
  bibtex_show = {true},
  thumbnail = {/assets/img/tpami22_thumbnail.png}
}

Characterizing Generalization under Out-Of-Distribution Shifts in Deep Metric Learning

Milbich*, T., Roth*, K., Sinha, S., Schmidt, L., Ghassemi, M., and Ommer, B.

In Advances in Neural Information Processing Systems (NeurIPS) 2021

NeuRIPS

arXiv Bib Code

@inproceedings{milbich2020ooDML,
  title = {Characterizing Generalization under Out-Of-Distribution Shifts in Deep Metric Learning},
  author = {Milbich*, T. and Roth*, K. and Sinha, S. and Schmidt, L. and Ghassemi, M. and Ommer, B.},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2021},
  selected = {true},
  arxiv = {2107.09562},
  bibtex_show = {true},
  abbr = {NeuRIPS},
  code = {https://github.com/CompVis/Characterizing_Generalization_in_DML},
  thumbnail = {/assets/img/neurips21_thumbnail.png}
}

Revisiting Training Strategies and Generalization Performance in Deep Metric Learning

Roth*, K., Milbich*, T., Sinha, S., Gupta, P., Ommer, B., and Cohen, J. P.

In International Conference on Machine Learning (ICML) 2020

ICML

arXiv Bib Code

@inproceedings{roth2020revisiting,
  title = {Revisiting Training Strategies and Generalization Performance in Deep Metric Learning},
  author = {Roth*, K. and Milbich*, T. and Sinha, S. and Gupta, P. and Ommer, B. and Cohen, J. P.},
  year = {2020},
  booktitle = {International Conference on Machine Learning (ICML)},
  selected = {true},
  arxiv = {2002.08473},
  bibtex_show = {true},
  abbr = {ICML},
  code = {https://github.com/Confusezius/Revisiting_Deep_Metric_Learning_PyTorch},
  thumbnail = {/assets/img/icml20_thumbnail.png}
}

Unsupervised Part-based Disentangling of Object Shape and Appearance

Lorenz, D., Bereska, L., Milbich, T., and Ommer, B.

In Conference on Computer Vision and Pattern Recognition (CVPR) 2019

CVPR

arXiv Bib Code Website

@inproceedings{cvpr19_unsup_disentangle,
  title = {Unsupervised Part-based Disentangling of Object Shape and Appearance},
  author = {Lorenz, D. and Bereska, L. and Milbich, T. and Ommer, B.},
  booktitle = {Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2019},
  selected = {true},
  arxiv = {1903.06946},
  bibtex_show = {true},
  abbr = {CVPR},
  website = {https://compvis.github.io/unsupervised-disentangling/},
  code = {https://github.com/CompVis/unsupervised-disentangling},
  thumbnail = {/assets/img/cvpr19_thumbnail.png}
}

Unsupervised Video Understanding by Reconciliation of Posture Similarities

Milbich, T., Bautista, M., Sutter, E., and Ommer, B.

In International Conference on Computer Vision 2017

ICCV

arXiv Bib

@inproceedings{iccv17_unsup_video,
  title = {Unsupervised Video Understanding by Reconciliation of Posture Similarities},
  author = {Milbich, T. and Bautista, M. and Sutter, E. and Ommer, B.},
  booktitle = {International Conference on Computer Vision},
  year = {2017},
  selected = {true},
  arxiv = {1708.01191},
  bibtex_show = {true},
  abbr = {ICCV},
  thumbnail = {/assets/img/iccv17_thumbnail.png}
}

Behavior-Driven Synthesis of Human Dynamics

Milbich*, T., Blattmann*, A., Dorkenwald*, M., and Ommer, B.

In Conference on Computer Vision and Pattern Recognition (CVPR) 2021

CVPR

arXiv Bib Code Website

@inproceedings{milbich2021behavior,
  title = {Behavior-Driven Synthesis of Human Dynamics},
  author = {Milbich*, T. and Blattmann*, A. and Dorkenwald*, M. and Ommer, B.},
  booktitle = {Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2021},
  selected = {true},
  arxiv = {2103.04677},
  bibtex_show = {true},
  abbr = {CVPR},
  website = {https://compvis.github.io/behavior-driven-video-synthesis/},
  code = {https://github.com/CompVis/behavior-driven-video-synthesis},
  thumbnail = {/assets/img/cvpr21behav_thumbnail.png}
}