10 papers
Learn2Synth: Learning Optimal Data Synthesis Using Hypergradients for Brain Image Segmentation
Xiaoling Hu, Xiangrui Zeng, Oula Puonti +3
Domain randomization through synthesis is a powerful strategy to train networks that are unbiased with respect to the domain of the input images. Randomization allows networks to s…
Deep generative priors for 3D brain analysis
Ana Lawry Aguila, Dina Zemlyanker, You Cheng +6
Diffusion models have recently emerged as powerful generative models in medical imaging. However, it remains a major challenge to combine these data-driven models with domain knowl…
Reference-Free 3D Reconstruction of Brain Dissection Slabs via Learned Atlas Coordinates
Lin Tian, Jonathan Williams-Ramirez, Dina Zemlyanker +14
Correlation of neuropathology with MRI has the potential to transfer microscopic signatures of pathology to in vivo scans. There is increasing interest in building these correlatio…
Learning to Upscale 3D Segmentations in Neuroimaging
Xiaoling Hu, Peirong Liu, Dina Zemlyanker +3
Obtaining high-resolution (HR) segmentations from coarse annotations is a pervasive challenge in computer vision. Applications include inferring pixel-level segmentations from toke…
A Modality-agnostic Multi-task Foundation Model for Human Brain Imaging
Peirong Liu, Oula Puonti, Xiaoling Hu +5
Recent learning-based approaches have made astonishing advances in calibrated medical imaging like computerized tomography (CT), yet they struggle to generalize in uncalibrated mod…
Conditional diffusion models for guided anomaly detection in brain images using fluid-driven anomaly randomization
Ana Lawry Aguila, Peirong Liu, Oula Puonti +1
Supervised machine learning has enabled accurate pathology detection in brain MRI, but requires training data from diseased subjects that may not be readily available in some scena…