1 code implementation • 2 Feb 2022 • Max-Heinrich Laves, Malte Tölle, Alexander Schlaefer, Sandy Engelhardt
In POTOBIM, we optimize both the parameters of the prior distribution and the posterior temperature with respect to reconstruction accuracy using Bayesian optimization with Gaussian process regression.
no code implementations • 31 Jan 2022 • Marcel Bengs, Finn Behrendt, Max-Heinrich Laves, Julia Krüger, Roland Opfer, Alexander Schlaefer
We analyze the value of age information during training, as an additional anomaly score, and systematically study several architecture concepts.
no code implementations • 11 Jun 2021 • Max-Heinrich Laves, Malte Tölle, Alexander Schlaefer, Sandy Engelhardt
Bayesian methods feature useful properties for solving inverse problems, such as tomographic reconstruction.
1 code implementation • 26 Apr 2021 • Max-Heinrich Laves, Sontje Ihler, Jacob F. Fast, Lüder A. Kahrs, Tobias Ortmaier
We apply estimation of both aleatoric and epistemic uncertainty by variational Bayesian inference with Monte Carlo dropout to regression tasks and show that predictive uncertainty is systematically underestimated.
no code implementations • 1 Jan 2021 • Max-Heinrich Laves, Sontje Ihler, Karl-Philipp Kortmann, Tobias Ortmaier
Various metrics have recently been proposed to measure uncertainty calibration of deep models for classification.
1 code implementation • 20 Aug 2020 • Max-Heinrich Laves, Malte Tölle, Tobias Ortmaier
We use a randomly initialized convolutional network as parameterization of the reconstructed image and perform gradient descent to match the observation, which is known as deep image prior.
no code implementations • 9 Jul 2020 • Sontje Ihler, Max-Heinrich Laves, Tobias Ortmaier
Using flow estimations from teacher model FlowNet2 we specialize a fast student model FlowNet2S on the patient-specific domain.
no code implementations • 20 Jun 2020 • Max-Heinrich Laves, Sontje Ihler, Karl-Philipp Kortmann, Tobias Ortmaier
The model uncertainty obtained by variational Bayesian inference with Monte Carlo dropout is prone to miscalibration.
2 code implementations • MIDL 2019 • Max-Heinrich Laves, Sontje Ihler, Jacob F. Fast, Lüder A. Kahrs, Tobias Ortmaier
The consideration of predictive uncertainty in medical imaging with deep learning is of utmost importance.
1 code implementation • 30 Sep 2019 • Max-Heinrich Laves, Sontje Ihler, Karl-Philipp Kortmann, Tobias Ortmaier
Model uncertainty obtained by variational Bayesian inference with Monte Carlo dropout is prone to miscalibration.
no code implementations • 2 Aug 2019 • Max-Heinrich Laves, Sontje Ihler, Tobias Ortmaier
Our approach uses the idea of deep image priors to combine convolutional networks with conventional registration methods based on manually engineered priors.
1 code implementation • 2 Aug 2019 • Max-Heinrich Laves, Sontje Ihler, Tobias Ortmaier
We evaluate two different methods for the integration of prediction uncertainty into diagnostic image classifiers to increase patient safety in deep learning.
no code implementations • 23 Mar 2019 • Max-Heinrich Laves, Sontje Ihler, Lüder Alexander Kahrs, Tobias Ortmaier
Noise in speckle-prone optical coherence tomography tends to obfuscate important details necessary for medical diagnosis.
1 code implementation • 23 Mar 2019 • Max-Heinrich Laves, Sontje Ihler, Lüder A. Kahrs, Tobias Ortmaier
A recent study established a diagnostic tool based on convolutional neural networks (CNN), which was trained on a large database of retinal OCT images.
no code implementations • 19 Jan 2019 • Max-Heinrich Laves, Sarah Latus, Jan Bergmeier, Tobias Ortmaier, Lüder A. Kahrs, Alexander Schlaefer
The resulting volumentric images provide additional information on the shape of caveties in the bone structure, which will be useful for image-to-patient registration and to estimate the drill trajectory.
no code implementations • 26 Oct 2018 • Max-Heinrich Laves, Lüder A. Kahrs, Tobias Ortmaier
Subsequent to this, the projections are warped by predicted lateral flow and 1D depth flow is estimated.
1 code implementation • 16 Jul 2018 • Max-Heinrich Laves, Jens Bicker, Lüder A. Kahrs, Tobias Ortmaier
Methods Four machine learning based methods SegNet, UNet, ENet and ErfNet were trained with supervision on a novel 7-class dataset of the human larynx.