Fabian Isensee
@FabianIsensee1
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 💀!
Amazing work! Congratulations 🎉
#ICLR2024 oral 🎉 We reveal critical gaps between research and application in uncertainty estimation for image segmentation, hindering the real-life benefits of methods. AC: “We really need this paper, and we really need it to get attention” arxiv.org/abs/2401.08501 (1/6) 🧵⬇️
Our work on improving perovskite #photovoltaics⚡️ manufacturing through #XAI is finally out in Advanced Materials: bit.ly/3QpoQJ3 🎉 We also presented this work last week at the #NeurIPS2023 XAIA workshop as one of 4 selected orals out of 59 📊 🧵⬇️
See you at #MICCAI2023 :) Check out our poster on “Understanding Silent Failures in Medical Image Classification” (T-03-125, on Tuesday morning) where we shed light on the critical cases where both classifier *and* uncertainty estimation (UE) fail: arxiv.org/abs/2307.14729 🧵⬇️
Update: We integrated @MetaAI's Segment Anything Model into Napari for both 2D and 3D images and extended #SAM to semantic and instance segmentation! Works out-of-the-box for natural, medical and most other types of images. github.com/MIC-DKFZ/napar…
The KiTS challenge is back with an expanded training set of 489 cases, brand new test set of 110 cases, and the addition of a new contrast phase! Help us advance the field & improve patient outcomes. Download the data today & compete for the $5,000 prize! kits-challenge.org/kits23
Exciting news! @mic_dkfz’s new Segment Anything Model (SAM) plugin for Napari is out now! One-click segmentation of any object with @Meta AI's SAM system, plus extension to full semantic segmentation. Check out the plugin here: ➡ github.com/MIC-DKFZ/napar… #segmentation #Napari"
Awesome! You are a symbol of useful and high-performance code!🫡 Recently, I did a survey on the winning solutions in MICCAI 2022 segmentation challenges. nnUNet keeps winning these competitions. Really amazing!