Excited to share our new paper (CVPR 2026 🚀): "MuViT: Multi-Resolution Vision Transformers for Learning Across Scales in Microscopy" which enables local predictions to use global context.
Great work led by @albertdm.bsky.social & another fun collab w @gioelelamanno.bsky.social! @scadsai.bsky.social
It's ok I guess I'll click on some cars or bikes to get my open-source PRs merged now
We push the SOTA in bacteria tracking once more, the remaining errors are almost cut in half 🎉
This work was spearheaded by the super-talented Cyril Achard in @maweigert.bsky.social's group 👏
For more on transformer-based tracking with pre-trained features, check github.com/C-Achard/Tra...
🚀⭐️ Trackastra update 🚀⭐️
We just released a new model that uses pre-trained image features from Meta’s SAM2 model.
This improves performance on many datasets, most notably for bacteria tracking.
To give it a try, simply install with the new optional dep `etultra` github.com/weigertlab/t...
Yes, and there is already a pretrained model trained on all Cell Tracking Challenge data, 2d + 3d.
We moved the AI@MBL course "Deep Learning for Microscopy Image Analysis" to HHMI Janelia (@hhmijanelia.bsky.social).
Join us for two weeks of intense lectures, exercise, and hands-on project work!
Course dates: June 4-18 2026
Application by: January 15 2026
www.janelia.org/you-janelia/...
Thanks @jytinevez.bsky.social for making the nice TrackMate integration
Give it a try following the tutorial made by @jytinevez.bsky.social imagej.net/plugins/trac...
Trackastra is part of the brand-new TrackMate v8.
It is the first deep-learning-based ("AI") drop-in replacement for TrackMate's regular LAP-Track algorithm.
Tracks out of the box for many types of datasets (bacteria, fluorescent nuclei, phase contrast cell culture etc), zero parameters to tune 🤖
Super cool!!! So many methods integrated is 🔥
Preprint -
Excited to present WHOLISTIC, which extends the concept of whole-brain functional imaging to the entire body. Pioneering work by incredibly talented Virginia Ruetten @vmsruetten.bsky.social, this platform reveals whole-organism cellular dynamics in vivo.
www.biorxiv.org/content/10.1...
Many of you already heard about this new algo that's faster than Dijkstra ( arxiv.org/abs/2504.17033 ), here I collected 3 software implementations: 1. #python github.com/sidharthpuna... 2. #rust github.com/alphastrata/... 3. #rust + some explanations of the paper medium.com/%40adnanmaso...
So cool! Causal up until the penultimate layer, I bet the DiGradCam will look interesting. Also, as commonly known, if you can 3d print your neural network architecture you get to name it! GiuseppeNet ;) ?
I love that! A simple thing might be to spin out orthogonal to the major axis (of movement) in the few previous frames, and the sample could be rotated according to that. Not sure if our brains are ready to deal with that
Thanks @janfunkey.bsky.social and team for organizing this fun Hackathon for making open source software for science. Many of the other lit 🔥 cell tracking things we worked on are available at github.com/live-image-t....
Bonus: I added a 3D version, here is an example on the C. elegans dataset from the Cell Tracking Challenge.
With this visualisation one can intuitively compare different trackers
I originally made these when developing my tracking algorithm Trackastra www.ecva.net/papers/eccv_..., and have finally made them easy-to-produce for anyone at this year's @hhmijanelia.bsky.social Trackathon.
Cell tracking is never perfect, and it's important to understand the types of errors your solution contains.
Here is my stab at this: Divisualisations in @napari.org. github.com/bentaculum/d...
You spin tracks out upwards from the playing video. Green edges are correct, FP in magenta, FN in cyan.
Looking for a great PhD position?
I had the pleasure of doing my doctoral research in Martin’s group - wholeheartedly recommended :)
Spotiflow, our deep learning based spot detection method for microscopy, is now published in @natmethods.nature.com!
Since the pre-print, we have added many features, notably native 3D detection!
@maweigert.bsky.social @gioelelamanno.bsky.social @epfl-brainmind.bsky.social
Paper: rdcu.be/epIB7
(1/N)
Bonus: Recently, I also successfully finished my PhD at EPFL 🥳
It was as an incredible journey, thanks once more to everyone that helped making it a success :)
First post and little update:
I’m super excited to share that I moved to the Bay Area and joined CZI as a Senior Research Scientist to continue working on Imaging & AI.
