toshas (Anton Obukhov)

posted an update about 2 months ago

Post

1860

Another gem from our lab — DGInStyle! We use Stable Diffusion to generate semantic segmentation data for autonomous driving and train domain-generalizable networks.

📟 Website: https://dginstyle.github.io
🧾 Paper: https://arxiv.org/abs/2312.03048
🤗 Hugging Face Paper: DGInStyle: Domain-Generalizable Semantic Segmentation with Image Diffusion Models and Stylized Semantic Control (2312.03048)
🤗 Hugging Face Model: yurujaja/DGInStyle
🐙 Code: https://github.com/yurujaja/DGInStyle

In a nutshell, our pipeline overcomes the resolution loss of Stable Diffusion latent space and the style bias of ControlNet, as shown in the attached figures. This allows us to generate sufficiently high-quality pairs of images and semantic masks to train domain-generalizable semantic segmentation networks.

Team: Yuru Jia ( @yurujaja ), Lukas Hoyer, Shengyu Huang, Tianfu Wang ( @Tianfwang ), Luc Van Gool, Konrad Schindler, and Anton Obukhov ( @toshas ).

posted an update 2 months ago

Post

1922

Introducing Marigold-LCM 🌼 — a FAST version of the now popular state-of-the-art depth estimator! Thanks to the latent consistency distillation, it retains the precision of the original Marigold but reaches the solution in just a few steps!

Check out the teaser video attached below and play with the new demo - it accepts videos now! Also, meet the new team member: Tianfu Wang ( @Tianfwang )

🤗 Demo: prs-eth/marigold-lcm
🤗 Model: https://huggingface.co/prs-eth/marigold-lcm-v1-0
🤗 Original Marigold post: https://huggingface.co/posts/toshas/656973498012745
🤗 Paper: Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation (2312.02145)
🌐 Website: https://marigoldmonodepth.github.io
👾 Code: https://github.com/prs-eth/marigold
👾 Code: pip install diffusers

·

replied to their post 5 months ago

Updated to include the paper!

posted an update 5 months ago

Post

Introducing Marigold 🌼 - a universal monocular depth estimator, delivering incredibly sharp predictions in the wild! Based on Stable Diffusion, it is trained with synthetic depth data only and excels in zero-shot adaptation to real-world imagery. Check it out:

🤗 Hugging Face Space: https://huggingface.co/spaces/toshas/marigold
🤗 Hugging Face Model: https://huggingface.co/Bingxin/Marigold
🤗 Hugging Face Paper: Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation (2312.02145)
🌐 Website: https://marigoldmonodepth.github.io
👾 Code: https://github.com/prs-eth/marigold
👾 Code: pip install diffusers (check comments to this post for details!)
📄 Paper: https://arxiv.org/abs/2312.02145

Brought to you by the fantastic team from the Photogrammetry and Remote Sensing group of ETH Zurich: Bingxin Ke ( @Bingxin ), Anton Obukhov ( @toshas ), Shengyu Huang, Nando Metzger ( @nandometzger ), Rodrigo Caye Daudt, and Konrad Schindler.

·

Anton Obukhov