Björn Ommer

12

Papers

284

Total Citations

Papers (12)

ZigMa: A DiT-style Zigzag Mamba Diffusion Model

DepthFM: Fast Generative Monocular Depth Estimation with Flow Matching

Diff2Flow: Training Flow Matching Models via Diffusion Model Alignment

Latent Drifting in Diffusion Models for Counterfactual Medical Image Synthesis

SCFlow: Implicitly Learning Style and Content Disentanglement with Flow Models

TREAD: Token Routing for Efficient Architecture-agnostic Diffusion Training

Does VLM Classification Benefit from LLM Description Semantics?

Continuous, Subject-Specific Attribute Control in T2I Models by Identifying Semantic Directions

CAGE: Unsupervised Visual Composition and Animation for Controllable Video Generation

CleanDIFT: Diffusion Features without Noise

Stochastic Interpolants for Revealing Stylistic Flows across the History of Art

What If: Understanding Motion Through Sparse Interactions