Katja Schwarz

AI Research Scientist in Computer Vision / Machine Learning

prof_pic.png

I am a Research Scientist at Meta.

My research lies at the intersection of computer vision and graphics and focuses on 3D vision. In particular, I am interested in enabeling machines to infer 3D representations from sparse observations, such as 2D images. I am passionate about leveraging such representations for generative modeling in 2D and 3D.

I did my PhD in the Autonomous Vision Group (AVG) at Tübingen University and the Max Planck Institute for Intelligent Systems, advised by Andreas Geiger. Prior to this, I studied Physics at Heidelberg University where I received my bachelor degree in 2016 and master degree in 2018.

selected publications

  1. worlds.gif
    A Recipe for Generating 3D Worlds From a Single Image
    Katja Schwarz, Denis Rozumny, Samuel Rota Bulo, and 2 more authors
    In arXiv.org, 2025
  2. ggs.gif
    Generative Gaussian Splatting: Generating 3D Scenes with Video Diffusion Priors
    Katja Schwarz, Norman Müller, and Peter Kontschieder
    In arXiv.org, 2025
  3. multidiff.gif
    MultiDiff: Consistent Novel View Synthesis from a Single Image
    Norman Müller, Katja Schwarz, Barbara Rössle, and 4 more authors
    In CVPR, 2024
  4. wildfusion.gif
    WildFusion: Learning 3D-Aware Latent Diffusion Models in View Space
    Katja Schwarz, Seung Wook Kim, Jun Gao, and 3 more authors
    In International Conference on Learning Representations, 2024
  5. voxgraf.gif
    VoxGRAF: Fast 3D-Aware Image Synthesis with Sparse Voxel Grids
    Katja Schwarz, Axel Sauer, Michael Niemeyer, and 2 more authors
    In Advances in Neural Information Processing Systems (NeurIPS), 2022
  6. styleganxl.gif
    StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasets
    Axel Sauer, Katja Schwarz, and Andreas Geiger
    In ACM Trans. on Graphics, 2022
  7. frequency_bias.gif
    On the Frequency Bias of Generative Models
    Katja Schwarz, Yiyi Liao, and Andreas Geiger
    In Advances in Neural Information Processing Systems (NeurIPS), 2021
  8. graf.gif
    GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis
    Katja Schwarz, Yiyi Liao, Michael Niemeyer, and 1 more author
    In Advances in Neural Information Processing Systems (NeurIPS), 2020