Martin Segeroth

Shipped Statsmed v0.0.18 - Quick medical data analysis + figures. New/updated examples: ROC/AUC, Boxplots, Bland–Altman, Correlations. Try it: github.com/segemart/stats… #Python #DataScience #MedStats

Try yourself: github.com/wasserth/Total… Changelog: github.com/wasserth/Total…

Massive update for TotalSegmentator! We doubled MR training cases & added new structures: MR: - Individual vertebrae - Appendicular bones - Body - Thigh+shoulder muscles (also for CT) CT: - Intermuscular fat - Lung nodules (thx to Bluemind AI) - Oculomotor muscles

4. awesome work with Lorraine Abel, @JakobWasserthal, Manfred Meyer, @Voss_MD, Shan Yang, Ricardo Donners, @Doc_X_Ray, Daniel Boll, Elmar Merkle, Christian Breit

3. We used volume deviation as a metric, with a 5% deviation considered reproducible. Non-inferiority testing was applied to assess this. - TotalSegmentator outperformed the nnU-Net especially in cases with pathologic findings.

Ever wondered how to assess reproducibility of an segmentation tool for medical images? Check out our latest paper on intra-individual reproducibility of automated abdominal organ segmentation: rdcu.be/dUoy3

5) The idea is to keep Statsmed simple and we will implement additional functions and create new examples. If you have ideas for improvement, please contact me.

4) We also implemented a functional t-test, that allows you to compare two groups of functions and was used in the following research article (sciencedirect.com/science/articl…)

3) Statsmed allows you to make figures quickly, see github.com/segemart/stats… for Bland-Altman plots

2) There are already some examples for you: Ex1 - Test for normal distribution Ex2 - Descriptive Statistic Ex3 - Pearson's and Spearman's correlation with figures Ex4 - Comparison of two groups with continuous variables Ex5 - Bland-Altman plots

1) Let me introduce you to Statsmed (github.com/segemart/stats…) - We designed Statsmed for you to perform statistical analysis of medical data easily and get a fast overview of you data with python

We added support for MR images to TotalSegmentator. It now segments 59 structures in MR images, regardless of sequence and contrast. code: github.com/wasserth/Total… paper: arxiv.org/abs/2405.19492 annotated dataset: doi.org/10.5281/zenodo… online tool: totalsegmentator.com

Is the size of an organ within the norm? Our free online tool automatically calculates organ volumes from CT scans and compares them to the standard population distribution, providing a percentile ranking. Try it out: compute.totalsegmentator.com/volume-report/

Incredible teamworkwith @Voss_MD, @JakobWasserthal, Hanns-Christian Breit and Tobias Heye

1/2 Ever wondered about the connection between polytrauma-CT and weather? Check out our latest article: Radiology Weather Forecast: A Retrospective Analysis of predictability of Median Daily Polytrauma-CT Occurrence Based on Weather Data. Read more: doi.org/10.1016/j.ejra…

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…

TotalSegmentator provides automatic, easily accessible segmentations of major anatomic structures in CT images buff.ly/472t4h8 @JakobWasserthal @MSegeroth #DeepLearning #MachineLearning #AI

4/4 💻 Code + Model: Run it all on your laptop! Check it out on GitHub: github.com/wasserth/Total… 📂 Dataset: Open access! Explore the annotated dataset here: zenodo.org/record/6802614 🔧 Online Tool: For quick and easy segmentation, try our online tool: totalsegmentator.com

3/4 Age is not just a number! Our research explored age-dependent volume & attenuation changes in various organ groups using 4004 whole-body CT exams. Fascinating correlations found! Check out the supplement.

1/4 Exciting news! 🚀 Totalsegmentator has officially been published in Radiology: Artificial Intelligence (pubs.rsna.org/doi/10.1148/ry…). Gratitude to our exceptional team for their hard work and dedication @JakobWasserthal @_JoshyCyriac_ @MPradella + Co-Authors! 👏