Björn Ommer

Adapting Self-Supervised Representations as a Latent Space for Efficient Generation

Purrception: Variational Flow Matching for Vector-Quantized Image Generation

Inferring Attribute Subspaces from Visual Contexts

Uncertainty Preservation in Generative Visual Autoregression

DisMo: Disentangled Motion Representations for Open-World Motion Transfer

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