torchio: Medical imaging toolkit for deep learning Lang: Python ⭐️ 1754 Author: Fernando Pérez-García #MachineLearning github.com/fepegar/torchio

We have two postdoc positions open on machine learning in medical imaging Imperial EEE I-X and BioMedIA Imperial, working on exciting topics of robust and trustworthy MRI! Please get in touch if you are interested. Please find more details here: cq615.github.io/vacancies/

Our TotalSegmentator paper was accepted at Radiology: Artificial Intelligence paper: pubs.rsna.org/doi/10.1148/ry… open preprint: arxiv.org/abs/2208.05868 code+model: github.com/wasserth/Total… online tool: totalsegmentator.com dataset: doi.org/10.5281/zenodo…

Check out Open Health Imaging Foundation running on Hugging Face Spaces, thanks to Radamés Ajna! 😍 huggingface.co/spaces/radames…

TotalSegmentator v2 is here: - 33 new classes - up to 32x faster runtime on CPU - improved label quality - larger training dataset (n=1559) - uses nnU-Net v2 Details: github.com/wasserth/Total…

Potential risks for #Foundation models in medical imaging #AI doi.org/10.1148/ryai.2… Charles Jones @mel_bernhardt Imperial College London #ChestRad #FoundationModels #ML

GPT4 and images. Exciting times, but still some work to be done. Be careful about how you name your files. 😉

We're delighted to bring #MICCAI2024 to Africa, for the very first time! Join us in beautiful Marrakesh for an unforgettable learning and networking experience! To get a flavour of what to expect, check out this short introductory video drive.google.com/file/d/1gZL5A3… and stay tuned!

There is a saying in #radiology, “One view is no view.” When generating radiology reports for most exam types, one image is no image. #GPT4V is impressive, but you must ask the right questions & be careful not to overhype #AI, esp. in #healthcare. #GenAI #RadTwitter #MedTwitter

$3,200 as the publication fee is still a lot! Many folks can't really afford this and we believe such high fees are unjustified. At MELBA, our goal is to bring this down to ZERO. Currently, we charge $10 as a submission fee and that's it. And we are fully open access. Consider

Radon Transform (RT) was formulated in 1917 but remained useless in practice until CT scanners were invented in the 60s But RT isn't just for CTs. You can use it for measuring motion! RT g(p,θ): Shoot rays at θ+90 & offset p, measure line integrals of f(x,y) along the ray 1/6

💻 Do you write software for researchers? Try out the new #FAIR Software Checklist, a tool by the ARDC & Netherlands eScience Center @akademienl.social to assess how your software is aligned w/ the FAIR principles 👉 ardc.edu.au/article/new-se… #FAIRsoftware #ResearchSoftwareAU #RSE #BetterSoftwareBetterResearch

🚀 Introducing the Stanford AIMI Dataset Index – a diverse inventory of AI-ready datasets for machine learning in healthcare! We hope this resource will help democratize access to health data for AI development. Explore & contribute here: bit.ly/3RmEGFL #AIMIDatasetIndex

Excited to share our #NeurIPS paper analyzing the good, the bad and the ugly sides of data augmentation (DA)! DA is crucial for computer vision but can introduce class-level performance disparities. We explain and address these negative effects in: openreview.net/pdf?id=yageaKl… 1/9

EfficientSAM was just released and it's fast! 💨 With 20x fewer params, it is now 20x faster than the original SAM segmentation model, while staying in the same accuracy range. See below for the project page and an interactive Hugging Face space to try it out! ⬇️⬇️

No autonomous surgical robots without synthetic data! Our latest paper introduces SARAMIS, the first large-scale dataset capturing human anatomy as 3D rendering assets for robotic assisted and minimally invasive surgery, to be presented at NeurIPS Conference shorturl.at/qTUY1 🧵

PyTorch's design origins, its connection to Lua, its intertwined deep connection to JAX, its symbiotic connection to Chainer The groundwork for PyTorch originally started in early 2016, online, among a band of Torch7's contributors. Torch7 (~2010-2017) These days, we also

🍻EXCITING DAY: Our Nature Methods sister publications on #MetricsReloaded are out🎉. Recommendations: rdcu.be/dysg5 Pitfalls: rdcu.be/dysgX Helmholtz Imaging MICCAI SIG for Challenges ProjectMONAI NCT Heidelberg DKFZ Important for the community: 1/

You can now slice TorchIO images and subjects! >>> import torchio as tio >>> subject = tio.datasets.FPG() >>> subject[50:200, 50:-30].plot() The image origin in the spatial metatada is updated accordingly as the Crop transform is used under the hood. (orientation here is PIR+)

We are announcing that PyTorch will stop publishing Anaconda packages on PyTorch’s official anaconda channels. For more information, please refer to the following post on dev-discuss: dev-discuss.pytorch.org/t/pytorch-depr…