Prof. Dr. Björn Ommer
Chair of Machine Vision and Learning, Computer Vision & Learning Group | Ludwig-Maximilians-Universität in Munich
Chair of AI for Computer Vision and Digital Humanities / the Arts, Ludwig-Maximilians University Munich
Publications
-
DepthFM: Fast Monocular Depth Estimation with Flow Matching
Conference paperProject: KLIMA-MEMES -
Distillation of Diffusion Features for Semantic Correspondence
Conference paperProject: KLIMA-MEMES -
Cross-Image-Attention for Conditional Embeddings in Deep Metric Learning
Conference paperhttps://openaccess.thecvf.com/content/CVPR2023/papers/Kotovenko_Cross-Image-Attention_for_Conditional_Embeddings_in_Deep_Metric_Learning_CVPR_2023_paper.pdfProject: KLIMA-MEMES -
Diff2Flow: Training Flow Matching Models via Diffusion Model Alignment
Conference paperhttps://doi.org/10.48550/arXiv.2506.02221Project: KLIMA-MEMES -
Distillation of Diffusion Features for Semantic Correspondence
DistillDIFT
Conference paperhttps://arxiv.org/abs/2412.03512Project: KLIMA-MEMES -
Does VLM Classification Benefit from LLM Description Semantics?
Conference paperhttps://doi.org/10.48550/arXiv.2412.03512Project: KLIMA-MEMES -
Fast Monocular Depth Estimation with Flow Matching
DepthFM
Conference paperhttps://doi.org/10.48550/arXiv.2403.13788Project: KLIMA-MEMES -
Training Flow Matching Models via Diffusion Model Alignment
Diff2Flow
Conference paperhttps://doi.org/10.48550/arXiv.2506.02221Project: KLIMA-MEMES -
Continuous, Subject-Specific Attribute Control in T2I Models by Identifying Semantic Directions
Conference paperhttps://doi.org/10.48550/arXiv.2403.17064Project: KLIMA-MEMES -
Diffusion Features without Noise
CleanDIFT
Conference paperhttps://doi.org/10.48550/arXiv.2412.03439Project: KLIMA-MEMES -
Token Routing for Efficient Architecture-agnostic Diffusion Training
TREAD
Conference paperhttps://doi.org/10.48550/arXiv.2501.04765Project: KLIMA-MEMES -
Understanding Motion Through Sparse Interactions
What If
Conference paperhttps://doi.org/10.48550/arXiv.2510.12777Project: KLIMA-MEMES -
Stochastic Interpolants for Revealing Stylistic Flows across the History of Art
Conference paperhttps://openaccess.thecvf.com/content/ICCV2025/html/Ma_Stochastic_Interpolants_for_Revealing_Stylistic_Flows_across_the_History_of_ICCV_2025_paper.htmlProject: KLIMA-MEMES -
Implicitly Learning Style and Content Disentanglement with Flow Models
SCFlow
Conference paperhttps://doi.org/10.48550/arXiv.2508.03402Project: KLIMA-MEMES -
Governing AI innovation for good
VortragProject: KLIMA-MEMES -
Diff2Flow
Training Flow Matching Models via Diffusion Model Alignment
VortragProject: KLIMA-MEMES -
Generative KI und die Zukunft der Intelligenz
VortragProject: KLIMA-MEMES -
Generative KI und die Zukunft von Intelligenz
VortragProject: KLIMA-MEMES


