PeopleofSuperintelligence

Researcher

Dominik Lorenz

Dominik Lorenz researches generative computer vision, with work on image synthesis, video generation, and flow matching. He coauthored latent diffusion research linked to Stable Diffusion, whose original release he helped lead as an academic. A rectified-flow transformer paper he coauthored won an ICML 2024 Best Paper Award, and later work explored in-context image editing. As of October 2026, his research affiliation is Black Forest Labs, where he coauthored Self-Flow for image, video, and audio generation.

Timeline

  1. 2019

    Separating shape from appearance

    His unsupervised method learns object parts consistently to separate appearance from shape.

  2. 2021–2022

    Moving diffusion into latent space

    With Robin Rombach, Andreas Blattmann, Patrick Esser, and Björn Ommer, he coauthors High-Resolution Image Synthesis with Latent Diffusion Models at LMU Munich and Heidelberg University. The CVPR 2022 paper applies diffusion in pretrained autoencoders’ latent space to reduce computational requirements.

  3. 2022

    Leading the Stable Diffusion release

    He helps lead the original Stable Diffusion release as an academic, alongside Robin Rombach and Andreas Blattmann.

  4. 2023

    Extending generation to video

    He coauthors Stable Video Diffusion, a latent video model for high-resolution text-to-video and image-to-video generation.

  5. 2024

    Scaling text-to-image transformers

    He and his coauthors present a rectified-flow transformer architecture for high-resolution text-to-image synthesis. The paper receives an ICML 2024 Best Paper Award.

  6. 2025

    Editing images in context

    He coauthors FLUX.1 Kontext, a flow-matching model that combines image generation and editing with text and image context.

  7. As of October 2026

    Researching multi-modal flow

    His research affiliation is Black Forest Labs. In 2026, he coauthored Self-Flow, a self-supervised flow-matching method that supports image, video, and audio generation.

Sources and credits

Photo: We have no published photo of Dominik Lorenz yet. Send us a free-licensed photo.

  1. Unsupervised Part-Based Disentangling of Object Shape and Appearance arxiv.org. Checked 14 October 2026.
  2. High-Resolution Image Synthesis with Latent Diffusion Models arxiv.org. Checked 14 October 2026.
  3. High-Resolution Image Synthesis with Latent Diffusion Models — Computer Vision & Learning Group ommer-lab.com. Checked 14 October 2026.
  4. Celebrating one year(ish) of Stable Diffusion … and what a year it’s been! — Stability AI stability.ai. Checked 14 October 2026.
  5. Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasets arxiv.org. Checked 14 October 2026.
  6. Scaling Rectified Flow Transformers for High-Resolution Image Synthesis arxiv.org. Checked 14 October 2026.
  7. Axel Sauer axelsauer.com. Checked 14 October 2026.
  8. FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space arxiv.org. Checked 14 October 2026.
  9. Self-Supervised Flow Matching for Scalable Multi-Modal Synthesis — Black Forest Labs bfl.ai. Checked 14 October 2026.
  10. Self-Supervised Flow Matching for Scalable Multi-Modal Synthesis arxiv.org. Checked 14 October 2026.