Stefan Winzeck
@SWinzeck
Followers
119
Following
297
Media
1
Statuses
70
he/him - post-doctoral Cambridge Residency @MSFTResearch, previously: post-doc @BioMedIAICL, PhD @Cambridge_Uni, MSc @ImperialCollege & @TU_Muenchen
Joined July 2020
.
Only 3% of Black students learn computer science in high school or beyond. Please watch and share this video. Inspire a student. Together we can change the face of computer science. #BlackVoicesForCS
0
0
0
In image analysis, we often assume subjects may be easily identified as subgroup members. But what if this information is only partially encoded in the data? Our MICCAI 2023 paper investigates "The Role of Subgroup Separability in Group-Fair Medical Image Classification". (1/6)
1
2
11
Network neuroscience is awesome ! 🧠But there are soo many ways to go from fMRI data to a connectome that you can analyse... 😵 Are they all just as good? Not at all! We compare 576 unique pipelines on 3 datasets across 5 different criteria, so you don't have to!
Interested in applying graph theory to your fMRI data but not sure how to best construct a network? We systematically compared network construction pipelines to arrive at methodological recommendations: https://t.co/CMD0gGnnpk with @loopyluppi @estamatakis Stavros Dimitriadis
0
40
188
Great team effort! Thanks to all the developers, radiology readers and trials unit team!! ML reduced round 2 reading-time by 11% with equivalent diagnostic performance @profstuartaylor @GlockerBen @ProfVickyGoh @BharwaniNishat @NIHR @ImperialMed
Development and Evaluation of Machine Learning in Whole-Body Magnetic Resonance Imaging for Detecting Metastases in Patients With Lung or Colon Cancer: A Diagnostic Test Accuracy Study. URL:
0
5
17
Join us through the fascinating world of brain connectivity at @CD_MRI! 🧠🔗🔬 Hear from our keynotes: Dr. Chun-Hung Yeh 🇹🇼, Dr. Demian Wassermann 🇫🇷, and Dr. Jennifer A. McNab 🇺🇸Don't miss this unique opportunity to explore the latest advancements! #MICCAI2023 @MICCAI_Society
1
10
29
Very glad to share our paper was accepted at the Journal of Neurotrauma. Check out the paper published in collaboration with @CenterTBI @SWinzeck @jansijbers @vfjn2 @Menon_Cambridge
liebertpub.com
The prediction of functional outcome after mild traumatic brain injury (mTBI) is challenging. Conventional magnetic resonance imaging (MRI) does not do a good job of explaining the variance in...
0
3
10
We are at an inflection point for the use of AI in health care. Microsoft Research’s Health Futures organization shares its work to make health care more data-driven, predictive, and precise:
3
16
48
Editor pick from March issue @GlockerBen & coll shed some light on how dataset biases manifest in predictive models by exploring methodology for subgroup analysis in image-based disease detection models. Read https://t.co/24UbBYmPna
#bias #disparities #fairness
0
2
8
In semantic segmentation, background samples provide key contextual information for segmenting regions of interest (ROIs). Our lastest @IEEE_TMI paper systematically investigates the effect of context labels on model learning, especially with small ROIs. https://t.co/Ztrsy6jsrd
7
9
61
Stroke multi-centre study driven by @mreyesag and colleagues. https://t.co/KNUnIvx97I
0
2
13
Exciting new work by @a_bourached while visiting the #RostLab: In what situation may #deep #learning enhance the prediction of acute #stroke #severity? 👉 https://t.co/N0Lyt8yYJa
#ResilientBrainLab #MRIGENIE @MGHNeurology @UCLIoN @nsanar 🥼🖥️🧠
0
5
14
Do you provide #CriticalCare to patients with #TBI ? We found that serum biomarkers could reduce the need for #MRI by 20-30% whilst being cost-saving Find out more in our latest paper https://t.co/wjZDCvNIrc
@Crit_Care Thanks @CenterTBI @Menon_Cambridge @vfjn2 & co-authors
0
7
18
ARE YOU POOLING multi-site neuroimaging data? Yes, you! ;) Do you the differences between ComBat🤺, Hierarchical Bayes regression, normative methods, + CycleGANs🚴? Johanna Bayer's guided tour explains when, why, + how to use these methods https://t.co/Ja7Ppfllfo
0
41
119
Hello neuroimagers! @CenterTBI @enigmabrains @ADNI3study Do you use multi-site #MRI data & wonder what the best ComBat harmonization method is for your data? Look @ our new paper https://t.co/e7TvOBgH2Q Building on great work of @chrisdav66 @takishinohara et al Thx @vfjn2
1
7
30
The latest HUMAP version is out! Now, it supports reproducible results. pip install humap==0.2.1
2
21
189
Many MNI spaces! Also many parcellations/templates in many different spaces. This is one of my favorite resources from Lawrence et al. and JHU team: https://t.co/vXaVFnJYb7. Huge number of parcellations all in MNI152NLin6 space.
One painful lesson I learnt in brain imaging: there are many MNI spaces, and yours is likely not the same as mine.
5
27
115
Want to run connectomics analyses🧐even in absence of DWI data? We provide a new multi-scale probabilistic atlas of the human connectome and user-friendly code to match it to your data🧠led by Yasser Aléman-Gómez @MeriBach @maxdescoteaux @pahagman et al.
1
36
118
how star wars characters would react if you asked them to buy you pads- a thread
90
1K
6K
Labeling bio images for AI just got 10x faster🏎. Today, we’re releasing Biodock Autolabel, and it’s kind of magical. We trained it on thousands of bio images so you can use it without training. (1/6) @BiodockAI #microscopy #pathology #biology
36
421
2K
@OdedRechavi Cold emailed a prof telling him "I will kill bacteria in the name of our Lord Darwin". Next thing I know, I have to rush to move to another country to start a PhD. From nerd to formal nerd. Still have the email.
0
2
35