Very happy that MuViT, our (CVPR 2026!) work on learning across spatial scales in microscopy w transformers, is out! Check Martin's thread for a nice walkthrough :)
Big thanks to my advisors @maweigert.bsky.social @gioelelamanno.bsky.social!
π₯οΈ: github.com/weigertlab/muvit
π: arxiv.org/abs/2602.24222
It was super cool to hear @alex-krull.bsky.social talking about his research at @humantechnopole.bsky.social. Alex never fails to inspire his audience π« π€©
Great work by our colleagues
@afoix.bsky.social
@virginieuhlmann.bsky.social
@alex-krull.bsky.social
As presented at this yearβs @neuripsconf.bsky.social
If you are at @neuripsconf.bsky.social and interested in shapes, check out Anna's work! Thank you @afoix.bsky.social! It was a real pleasure working on this with you and @virginieuhlmann.bsky.social and I am sad I can't join in San Diego.
Weβll be at @neuripsconf.bsky.social on Wednesday, 3 December, from 11 a.m. to 2 p.m. (San Diego time). Come and chat with us about why distance matrices and invariances are cool! π #InterdisciplinaryML #neurips2025 #bioimageanalysis #shapequantification
Cheers, @afoix.bsky.social. Thank you all of you for submitting, reviewing, organising and attending! Thank you #ICCV2025 for hosting us. Also thank you @unibirmingham.bsky.social for sponsoring #BIC again this year!
Thank you, @afoix.bsky.social!
I am really happy I could be part of this one. Everybody, please enjoy the new ShapeEmbeding goodness!
Thank you @afoix.bsky.social , thank you @virginieuhlmann.bsky.social! I am very happy and proud I was part of this one!
Cheers, for sharing, @paveltomancak.bsky.social! What a brilliant example for rotation-invariance or the lack of it.
Will go into the lecture.
Sept 18: Learn how to remove noise from imaging data with supervised, self-supervised & generative AI. With @alex-krull.bsky.social (University Birmingham).
Register for the series π bit.ly/6-image-proc...
@helmholtz.de
#imaging #ImageDenoising #AI
Looking forward to my talk at heidelberg.ai. I will touch on topics like #MapFreeReloc and #ACEZero, i.e. how to derive scene representations from very few (one!) or very many images.
Please meet Dr. @stonks1.bsky.social!
Couldnβt be more proud!
Submit your method to our denoising challenge!!! π
Do you like developing new AI vision methods for microscopy image analysis? You love theory & implementation? 1 week left to apply for a fully funded PhD position in our lab in Dresden π©πͺ! Topics: object detection/tracking, multimodal models & more. DM/email for details! #PhD #AcademicJobs #GPUsgoBrr
Rock on, Pavel! πππ
The 2nd AI4Life Challenge is live!
Calling the AI & bioimaging community to tackle a key microscopy challenge: removing noise while preserving detail.
π¦ Paired noisy/clean datasets
π Ground-truth evaluation
π§ DL focus
Build, test, compete π ai4life.eurobioimaging.eu/challenge-2/
Join the challenge, get the data, contribute your solution!!! π
The submission for 2025's BioImage Computing @iccv.bsky.social is now open. If things go like the last years, I am looking forward to all you brilliant submissions.
www.bioimagecomputing.com
Congrats, guys!
Improvements in LM (top) and EM (bottom) of our micro-sam model (finetuned) compared to the default SAM model.
After a long journey, Segment Anything for Microscopy is now published in Nature Methods! We significantly improve SAM for interactive and automatic segmentation in light and electron microscopy and build a user-friendly tool.
www.nature.com/articles/s41...
Tweedie's formula is super important in diffusion models & is also one of the cornerstones of empirical Bayes methods.
Given how easy it is to derive, it's surprising how recently it was discovered ('50s). It was published a while later when Tweedie wrote Stein about it
1/n
"Having trouble selling your #Tesla? Tired of being associated with a toxic, Nazi-adjacent brand? We understand that thereβs a shrinking market of buyers and limited selling options for current owners.
Our experienced team specializes in discreetly acquiring and transporting your car to [β¦]
Wow! Congrats!
@docmilanfar.bsky.social , love this. Max likelihood points to x. This refers to the fact that the noisy data x is actually the max likelihood est. of u?
Have you said thank you once to Tweedie?
π¨ Fantastic opportunity to conduct a PhD in Europe! European funding and fantastic international community! π©βπ
Poster for Theory of Living Matter seminar series talk 05 March 2025 at 16:00 UK time. Speaker is Virginie Uhlmann from BioVisionCenter, University of Zurich and EMBL-EBI. The talk is entitled "Turning morphology into numbers across microscopy scales and modalities", with the following abstract: Microscopy image data constitute the primary source of observational data on living systems across scales, and recent advances in microscopy technologies now enable the multi-modal observation of identical biological specimen. A key question emerges from this abundance of image data: how can we consistently quantify what we see? This endeavour is particularly complex for morphology, which involves intricate combinations of visual features that appear fundamentally different across various imaging modalities. Motivated by this challenge, my research group has been focusing on abstracting the problem of image-based morphology quantification from specific imaging modalities. In this seminar, I will present how we are building bits by bits the computational tools needed to extract modality-agnostic shape descriptors from microscopy images, how we use them to investigate fundamental biological questions, and what are the promising directions we are pursuing towards establishing a general framework to quantify biological morphology across imaging modalities.
π¨Next week, Wed 05/03 | 16:00 UK - Online Seminarπ¨
"Turning morphology into numbers across microscopy scales and modalities" by Dr Virginie Uhlmann.
To attend online please register to our π§ for the zoom link: lists.cam.ac.uk/sympa/subscr...
Pflanzenzellen, die mit einem Fluoreszenzmikroskop aufgenommen und mit dem Modell automatisch segmentiert wurden. Die zugrunde liegenden Daten sind dreidimensional und das Bild zeigt eine Darstellung der segmentierten Zellen, die jeweils durch eine andere Farbe reprΓ€sentiert werden. Foto: Nature Methods: 10.1038/s41592-024-02580-4
Segmentierung von Zellen in der Lichtmikroskopie mit ΞΌSAM. Das Bild zeigt, wie Zellen in der Phasenkontrastmikroskopie mit ΞΌSAM segmentiert werden kΓΆnnen. GrΓΌne Punkte und KΓ€stchen zeigen die Benutzereingabe und farbige Masken die entsprechende Vorhersage des Modells. Foto: Erstellt von Anwai Archit mit dem ΞΌSAM-Tool, verfΓΌgbar in Nature Methods: 10.1038/s41592-024-02580-4
Automatische Zellanalyse mit #KI: Forschende trainierten eine bestehende, KI-basierte Software neu. Das Modell βSegment Anything for Microscopyβ kann Bilder von Geweben, Zellen und anderen Strukturen genau segmentieren: s.gwdg.de/HqkMz2; s.gwdg.de/iTejKW
Forschungsteam mit bsky.app/profile/cppa...
Loved it!
