Geometric Deep Learning for the Assessment of Thrombosis Risk in the Left Atrial Appendage

no code implementations • 19 Oct 2022 • Xabier Morales, Jordi Mill, Guillem Simeon, Kristine A. Juhl, Ole De Backer, Rasmus R. Paulsen, Oscar Camara

The assessment of left atrial appendage (LAA) thrombogenesis has experienced major advances with the adoption of patient-specific computational fluid dynamics (CFD) simulations.

Was that so hard? Estimating human classification difficulty

no code implementations • 22 Mar 2022 • Morten Rieger Hannemose, Josefine Vilsbøll Sundgaard, Niels Kvorning Ternov, Rasmus R. Paulsen, Anders Nymark Christensen

In this paper, we introduce methods for estimating how hard it is for a doctor to diagnose a case represented by a medical image, both when ground truth difficulties are available for training, and when they are not.

Classification Metric Learning

Multi-modal data generation with a deep metric variational autoencoder

no code implementations • 7 Feb 2022 • Josefine Vilsbøll Sundgaard, Morten Rieger Hannemose, Søren Laugesen, Peter Bray, James Harte, Yosuke Kamide, Chiemi Tanaka, Rasmus R. Paulsen, Anders Nymark Christensen

We present a deep metric variational autoencoder for multi-modal data generation.

Data Augmentation

Multi-view consensus CNN for 3D facial landmark placement

1 code implementation • 14 Oct 2019 • Rasmus R. Paulsen, Kristine Aavild Juhl, Thilde Marie Haspang, Thomas Hansen, Melanie Ganz, Gudmundur Einarsson

The current state-of-the-art approaches have yet to gain from the dramatic increase in performance reported in human pose tracking and 2D facial landmark placement due to the use of deep convolutional neural networks (CNN).

Anatomy Pose Tracking