Uwe Schmidt

If you're interested in the details of our #StarDist-based submission to the 2022 CoNIC challenge, we finally released the code. github.com/stardist/stard…

@EmtiyazKhan Congrats!

I am pleased to be promoted to a "permanent PI" position at RIKEN (equivalent to a "tenured" full professor). Thanks to all my collaborators and mentors for their help, and all the funding agencies that supported our work. My position starts next month! 1/

@PavelTomancak @CEITEC_Brno @muni_cz

I am whole again. My little mobile coffee setup at the #AmazingApartments close to ⁦@CEITEC_Brno⁩ ⁦@muni_cz⁩. With #kofikofi beans => kickass cappuccino. Ready for more #LISH22

Together with the wonderful collaborators and community partners, we are happy to finally present the BioImage Model Zoo at bioimage.io, now described in biorxiv.org/content/10.110….

How well does #StarDist work on histopathology images? Pretty well! Our StarDist approach (with @uschmidt83) just won the nuclei seg/class task of the ISBI 2022 CoNIC challenge! 🎉 Paper: arxiv.org/abs/2203.02284 conic-challenge.grand-challenge.org/evaluation/seg…

I get a lot of reviews that say my work is not novel and I bet I'm not alone. It's always frustrating because I see novelty where the reviewer doesn't. Rather than rebut every critique, I've written a blog post to help reviewers think about novelty. perceiving-systems.blog/en/news/novelt…

Our new paper on "The Bayesian Learning Rule" is now on arXiv, where we provide a common learning-principle behind a variety of learning algorithms (optimization, deep learning, and graphical models). arxiv.org/abs/2107.04562 Guess what, the principle is Bayesian. A very long🧵

@psteinb_ @haesleinhuepf @matplotlib @helenajambor @annebonnerdata @HIDAdigital @thecarpentries @bercowskya @NEUBIAS_COST @bokeh The scikit-image part in my repo covers some plotting with matplotlib. But it’s all tailored to biologists. github.com/uschmidt83/neu…

@haesleinhuepf @matplotlib @helenajambor @annebonnerdata @HIDAdigital @thecarpentries @bercowskya @uschmidt83 @NEUBIAS_COST Lovely dive into plotting and later using @bokeh . Thanks for sharing.

I was just asked: "Can you recommend some online courses that teach me how to plot? I never did that." 🤔📉The @matplotlib tutorials appeared to me a useful start, but maybe there're alternatives for the #pydataverse? @helenajambor @annebonnerdata @HIDAdigital @thecarpentries

We just made a new stardist release, one of the bigger ones! Multi-class prediction, memory/runtime gains, and integration for the shiny new napari plugin :) As usual, great colab with @uschmidt83 and many others: github.com/stardist/stard…

@roybiolab @FijiSc I’m not sure, maybe related to this? forum.image.sc/t/warning-your… If that doesn’t help, please create a new topic at the image.sc forum.

@uschmidt83 @FijiSc was so excited, then got this error when I was trying to install the plugin. Have you seen this before?

We released a #StarDist plugin for @FijiSc! 🎉 imagej.net/StarDist It ships with a versatile model that we found to work well for segmenting cell nuclei in a variety of challenging (2D) fluorescence microscopy images. 🌟 #WeAreStarDist

@psobolewskiPhD @FijiSc Yes, tensorflow is the real issue when it comes to StarDist.

@uschmidt83 @FijiSc Alas, compiling Apple Silicon native TensorFlow bindings seems like a pretty big undertaking, bigger than my go at JOCL for sure...

@psobolewskiPhD @FijiSc I think Fiji can’t run native at the moment.

@uschmidt83 @FijiSc I tried that, with no luck—are you running native Fiji or in Rosetta?

@psobolewskiPhD @FijiSc forum.image.sc/t/fiji-crashin…

@uschmidt83 @FijiSc I'll try on Monday—can't get it running on my M1 MBPro.

@psobolewskiPhD @FijiSc I don’t really know. Have you tried running StarDist on a crop (1000x1000 or so) and does it work there as expected?

@uschmidt83 @FijiSc Hi! I was playing with your plugin on some really trivial data—tile scans of Hoechst stained nuclei. Segments nicely, but only the first 60 or so objects? Is this a file size/memory issue? (7xxx by 5xxx px) twitter.com/psobolewskiPhD…

@haesleinhuepf @napari_imaging Thanks! How did you find out about this feature? By the way, it works for me in standard Python, but not when I run my script via IPython (nothing shows up).

@uschmidt83 @napari_imaging No need to do complicated stuff, just use warnings.warn() 🙂 #L40" target="_blank" rel="nofollow noopener">github.com/clEsperanto/py…

I really like how @napari_imaging displays user warnings and error messages since recently. Who ever made that deserves a 🏆 for user-convenience. 🤩

@JwPylvanainen FYI to people only on Twitter: forum.image.sc/t/fiji-crashin…

Which TensorFlow version should I install on Apple M1 for Stardist segmentation to run on Fiji?

@ahklemm @knime @stelfrich @jan_eglinger Cute logo was made by @SuperResoluSian :)

#StarDist Segmentation in @knime! Highlighted just now by @stelfrich and @jan_eglinger at #NeubiasAcademy. Also wins the price for the cutest node🌟

@EntraCod There is a reason why the memory grows during training (cache, cf. #issuecomment-765795762" target="_blank" rel="nofollow noopener">github.com/stardist/stard…). We will add an option to disable this cache in the next release. In general, please discuss issues like this in the image.sc forum or open an issue.

@EntraCod Nobody prevents you extracting patches from your large labeled stacks upfront to reduce your training data. However, when using larger stacks for training, StarDist will dynamically sample different patches each time the stack is used.