An AI research lab within the Department of Radiology at Duke University.
Durham, NC
Joined December 2021
- Tweets273
- Following208
- Followers302
- Likes53
Mazurowski Lab - Duke University retweeted
I'm happy to share that our lab had three short papers and a long paper accepted at @midl_conference!
(1/4): Rethinking Perceptual Metrics for Medical Image Translation
(openreview.net/forum?id=acNu…), led by me;
Mazurowski Lab - Duke University retweeted
Excited to share our new paper, "ContourDiff: Unpaired Image Translation with Contour-Guided Diffusion Models”, which leverages domain-invariant anatomical contour representations to preserve anatomy through translation! (1/N)
arxiv.org/abs/2403.10786
Code: github.com/mazurowski-lab/Co…
Mazurowski Lab - Duke University retweeted
How and why do neural nets learn differently from natural images vs. medical images? I’m happy to announce my new #ICLR2024 paper, “The Effect of Intrinsic Dataset Properties on Generalization”! arxiv.org/abs/2401.08865
Check out the code: github.com/mazurowski-lab/in…
More info next..
Mazurowski Lab - Duke University retweeted
This article from @benWTmd, @MazurowskiPhD, et al. compared the diagnostic performance of ACR TI-RADS and a deep learning algorithm in differentiating benign and malignant thyroid nodules on ultrasound in children and young adults.
ajronline.org/doi/10.2214/AJ…
Mazurowski Lab - Duke University retweeted
Which examples in a neural net’s training data were important for learning interpretable concepts? Our paper “Attributing Learned Concepts in Neural Networks to Training Data” has been accepted for an oral presentation at the #NeurIPS2023 ATTRIB Workshop! arxiv.org/abs/2310.03149
Mazurowski Lab - Duke University retweeted
Excited to announce our new paper in IEEE Trans. in Med. Imaging, "SWSSL: Sliding window-based self-supervised learning for anomaly detection in high-resolution images", led by Haoyu Dong!
ieeexplore.ieee.org/abstract…
Code: github.com/mazurowski-lab/sw… (1/N)
Mazurowski Lab - Duke University retweeted
Happy to announce that our paper, "Segment Anything Model for medical image analysis: An experimental study" has just been published in Medical Image Analysis! sciencedirect.com/science/ar…
arXiv version: arxiv.org/abs/2304.10517
code: github.com/mazurowski-lab/se…
(1/2)
Mazurowski Lab - Duke University retweeted
The foreground-background imbalance problem can cause significant drops in object detection performance. Our new paper, “A systematic study of the foreground-background imbalance problem in deep learning for object detection”, is now on the arXiv! arxiv.org/abs/2306.16539 (1/n)
Mazurowski Lab - Duke University retweeted
New preprint from my lab: "Convolutional Neural Networks Rarely Learn Shape for Semantic Segmentation" by Yixin Zhang and @MazurowskiPhD! They study how segmentation CNNs learn and rely on object shape, the causes and applications of shape learning, etc. arxiv.org/abs/2305.06568
Mazurowski Lab - Duke University retweeted
Does Segment Anything Model (SAM) really work on medical images?
We tested it on 11 datasets and wrote a paper:
arxiv.org/abs/2304.10517
In brief: Performance varies widely across different datasets from impressive (given zero-shot setup) to poor.
Details follow. 🧵1/14.
Mazurowski Lab - Duke University retweeted
UPDATE: I have compiled all released lesion detection algorithm code into one convenient GitHub page here: github.com/mazurowski-lab/DB…
Mazurowski Lab - Duke University retweeted
We are excited to announce the March Virtual @duke_spark Seminar, by Prof. James S. Duncan of Yale University! The talk is entitled "Neuroimage Analysis in Autism: from Model-Based Estimation to Data-driven Learning". Registration and more info here: duke.zoom.us/webinar/registe…
Mazurowski Lab - Duke University retweeted
This paper by @MazurowskiPhD and colleagues @MikeHarowiczMD @Dr_Lars_Grimm @DrShelleyHwang evaluated whether deep-learning algorithms can accurately predict the presence of occult invasive disease for patients who were diagnosis with ductal carcinoma in situ. (1/3)
Mazurowski Lab - Duke University retweeted
This tutorial series by our Ph.D. student @KickNonz provides a beginner's introduction to doing medical image analysis with deep learning, and was just featured on Towards Data Science's Editors' Picks. Part 2: towardsdatascience.com/train…
Mazurowski Lab - Duke University retweeted
A common challenge in developing deep learning models to detect cancerous breast lesions is that only a limited number of scans contain cancerous lesions. For instance, less than 1% of digital breast tomosynthesis scans are found to contain cancer cases. (1/7)
Mazurowski Lab - Duke University retweeted
Join us for this seminar. I will talk about our public breast MRI dataset and about public sharing of data in general.
Join us for the second talk in the AI for breast MRI seminar, where we will be discussing the importance of data sharing for clinical benchmarks and common issues. This Thursday 29th, 11AM (ET) on Zoom:
sites.google.com/view/ai-bre…
Mazurowski Lab - Duke University retweeted
Check out our next seminar on medical imaging AI! Live on August 2nd at noon EST. Available to everybody! Register here: bit.ly/3ofcFko
Happening Tuesday, August 2nd at noon EDT: @DukeEngineering's Alberto Bartesaghi presents the latest entry in @DukeAIHealth’s monthly Spark Seminar series, hosted by @MazurowskiPhD: “Entering the Era of AI-driven Cryo-EM with SmartScope” Info/registration: bit.ly/3ofcFko
Mazurowski Lab - Duke University retweeted
Check out this new paper from @duke_spark's @MazurowskiPhD and the team! Just accepted to MICCAI 2022! The first author is @KickNonz from @MazurowskiLab .
I'm excited to announce that my paper "The Intrinsic Manifolds of Radiological Images and their Role in Deep Learning" (preprint here), was accepted to MICCAI 2022. Let me know if you'd like to discuss the paper, or if I'll be seeing you in Singapore! arxiv.org/abs/2207.02797
Mazurowski Lab - Duke University retweeted
In an article published recently in #JAMIA, a group of @DukeU authors describes a governance framework for deploying clinical #algorithms and risk prediction models: bit.ly/39OSS7J @DukeMedSchool @DukeInnovate @armandobedoyaMD #ABCDSOversight
Mazurowski Lab - Duke University retweeted
Duke AI Health’s Spark initiative (@duke_spark) convenes technical & clinical expertise across Duke to focus on the development, validation, and clinical implementation of #AI algorithms for applications related to medical imaging. aihealth.duke.edu/spark/
