Fabian Isensee

Excited to share nnInteractive: The nnU-Net moment for 3D Interactive Segmentation! Supports points, scribbles, boxes, & a new lasso prompt for full 3D masks from 2D inputs. Check it out: arxiv.org/abs/2503.08373
github.com/MIC-DKFZ/napar…
mitk.org/wiki/MITK-nnIn…

nnU-Net v2.6 released!
Most important change: Rework of the preprocessed data format 🚀 We now use Blosc2 for partially reading compressed files! Faster IO with less disk space required. Many thanks to Jan Sellner from @DKFZ_IMSY_lab and the @Blosc2 team for their amazing help!

HD-BET just became a lot more powerful and robust thanks to nnU-Net (+some data augmentation tweaks) 🥳
Install with `pip install -U hd-bet` or via our GitHub repository:
github.com/MIC-DKFZ/hd-bet
Happy brain extraction 💀!

📢 New nnU-Net variants 📢 The #MICCAI2024 challenge season has started and we are giving you the tools 🔧 to succeed. Use the new [and scalable] residual encoder presets in nnU-Net to get the most out of your datasets!

How do they perform? 👉 arxiv.org/pdf/2404.09556…

Concerning results: We've evaluated Transformer-, Mamba-, and CNN-based architectures, revealing widespread validation flaws in #MedicalImageSegmentation. Many claims of superiority do not withstand strict testing. Time to push for rigor in the field :) arxiv.org/pdf/2404.09556…