Enhancing Uncertainty Estimation in Semantic Segmentation via Monte-Carlo Frequency Dropout

John A. Onofrey; Lawrence H. Staib; Tal Zeevi

REVIEW 3 major objections 4 minor 28 references

Enhancing Uncertainty Estimation in Semantic Segmentation via Monte-Carlo Frequency Dropout

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Frequency dropout, applied during Monte-Carlo inference, produces uncertainty estimates that align more closely with where a pretrained segmentation model actually makes errors than standard signal dropout does, across MRI, CT, and X-ray…

desk verdict Useful segmentation extension of frequency dropout, but the central calibration advantage is unverified because the best configurations are chosen on the test set; worth a referee with a requested revision. read the letter →

arxiv 2501.11258 v1 pith:YHPHADDD submitted 2025-01-20 cs.CV cs.LGeess.IVstat.ML

Tal Zeevi , Lawrence H. Staib , John A. Onofrey This is my paper · ORCID

classification cs.CVcs.LGeess.IVstat.ML

keywords uncertaintyestimationsemanticsegmentationMonte-Carlodropoutfrequencycalibrationmedicalimagingselectiveprediction

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that dropout applied in the Fourier domain, rather than the pixel domain, produces uncertainty estimates that better reflect where a pretrained segmentation model actually fails. Using Monte-Carlo repetitions at inference time, the authors compare frequency-domain dropout with standard signal dropout across prostate MRI, liver CT, and lung X-ray segmentation. In their experiments, the best frequency-dropout configuration converged to better-calibrated uncertainty estimates, with standard deviations across repetitions aligning more closely with full-model segmentation errors. The authors also report that frequency dropout preserved segmentation accuracy, with Dice scores comparable to signal dropout and to the undiluted model.

What carries the argument

Frequency Dropout is the operative mechanism: during each forward pass, the feature map is transformed with the Fourier transform, multiplied element-wise by a binary mask drawn from a Bernoulli distribution, and transformed back before the nonlinearity. Because a single deleted Fourier coefficient affects the entire spatial domain, the induced noise is global and coherent rather than local impulse noise, which the paper argues preserves structural dependencies such as object boundaries.

What would settle it

Compute the inverse Fourier transform of a real feature map after applying the binary mask as defined in the paper, on any image from the Medical Segmentation Decathlon, and check whether the result has a non-negligible imaginary component; if it does, the forward pass as specified cannot produce real feature maps, and the method depends on an unstated symmetry constraint.

Watch

Extended reading notes

Core claim

The central claim is that stochastic attenuation of frequency components in CNN feature maps, sampled during Monte-Carlo inference, yields predictive uncertainty estimates that are better calibrated to true segmentation errors than the impulse noise introduced by standard signal dropout. The authors formalize frequency dropout as $\mathcal{C}^F_\theta(X) = \sigma(\mathcal{F}^{-1}(\mathcal{F}[X*W] \odot D) + b)$, where the binary mask $D$ deletes random Fourier coefficients. Across three modalities, the best-performing frequency-dropout configuration produced lower Expected Uncertainty Calibration Error and uncertainty maps whose spatial pattern matched the error regions of the full model. The method is presented as a drop-in inference-time procedure that works with pretrained models such as nnU-Net and MedSAM with no retraining.

Load-bearing premise

The forward pass only works as written if the real binary mask, when applied to the Fourier transform of a real feature map, preserves Hermitian symmetry so that the inverse transform returns a real signal; the paper does not state that the mask has this symmetry.

Editorial extensions

If this is right

MC-Frequency Dropout can produce uncertainty maps that better highlight voxels where a pretrained segmentation model errs, improving selective prediction and human review.
The method works without retraining, so it can be applied to any existing segmentation network first tested with signal dropout.
Best dropout placement is task-dependent: decoder dilution favored frequency dropout in liver and prostate, while encoder and global placements favored signal dropout in some settings.
Lower dropout rates generally yield better-calibrated uncertainty, except in liver tumor segmentation where higher rates were effective.
The extra FFT cost scales as $O(N \log N)$ per feature map, a trade-off that matters for large feature maps at inference time.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because the method's validity hinges on the Fourier mask preserving Hermitian symmetry, an implementation that enforces this constraint may be essential; if the released code does so, it should be documented.
The frequency-domain perturbation could be combined with test-time augmentation or with heteroscedastic noise to separate epistemic and aleatoric uncertainty.
The same principle might transfer to other global perturbations, such as wavelet-domain dropout, which could offer better spatial localization than Fourier modes.
A direct comparison of uncertainty quality against ensemble methods or MC-dropout with larger $R$ on the same pretrained models would clarify whether the calibration gain persists outside the chosen configurations.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: plausible, clearly written extension of frequency dropout to segmentation, with public code and three public datasets. But the headline claim is not actually established by the reported evidence, because the "best" dropout configuration for each method is selected on the same test set used for evaluation.

What's new: frequency dropout itself is not new, appearing as spectral dropout [11] and in the authors' own ISBI 2024 classification paper [10]. What is new is the systematic evaluation for segmentation: three modalities (MRI, CT, X-ray), three dropout layer placements, six rates, and an analysis of how placement interacts with calibration. That is a legitimate empirical contribution, and the paper is honest about the method's dependence on hyperparameters.

What's done well: public code, public datasets, sensible metrics (UCE for calibration, DSC for performance), and a fair comparison protocol with matched MC repetitions and rates. The discussion acknowledges limitations. No sign of overclaiming beyond a standard "potential to enhance" closing.

The main problem: Figure 2 compares the best-performing configuration per method, where "best" is found by searching 6 rates × 3 placements = 18 configurations, with UCE computed on the same small test cohorts (n=36, 37, 40). Picking the minimum UCE on the test set is a winner's-curse operation; two such minima can differ even when methods are equivalent. No validation split, no error bars, no multiple-comparison correction is reported. Since Figure 3 shows strong dependence on rate and placement, the selection step is exactly where bias enters. The central claim — that Frequency Dropout yields better-calibrated uncertainties — is therefore unverified as reported. This is more load-bearing than the Hermitian-symmetry question about Eq. (2), which is a real but minor implementation detail (using rfft/irfft with a symmetric mask fixes it; the paper just doesn't say).

There is also no significance testing for the UCE differences. The DSC table has Bonferroni-corrected p-values, but the calibration comparison does not. Minor, but worth noting.

Bottom line: the idea is worth testing and the paper deserves a serious referee, but as submitted the empirical advantage is not established. A revision with a held-out validation split for hyperparameter selection, error bars, and ideally a paired significance test would answer the main concern. I would send this to peer review rather than desk reject; it is a small conference paper with a fixable flaw.

Recommendation: engage with it, and ask for the validation split.

Referee Report

3 major / 4 minor

Summary. The paper proposes Monte-Carlo Frequency Dropout, a test-time stochastic perturbation method that applies dropout masks in the Fourier domain of feature maps, and evaluates it against standard MC Signal Dropout for uncertainty estimation in semantic segmentation. Experiments cover three public medical imaging tasks: prostate zone segmentation in biparametric MRI, liver and liver-tumor segmentation in contrast-enhanced CT, and lung segmentation in chest X-ray, using pretrained nnU-Net and MedSAM models. For each method, the authors sweep six dropout rates and three layer placements, run R=30 Monte-Carlo repetitions, and report uncertainty calibration error (UCE) and Dice divergence from the full no-dropout model. The headline claim is that the best-performing Frequency Dropout configuration achieves better-calibrated uncertainty estimates than the best-performing Signal Dropout configuration while preserving segmentation performance.

Significance. If the central empirical claim is established, MC-Frequency Dropout would be a simple, practical improvement for uncertainty quantification in medical image segmentation, with potential value for selective prediction and model scrutiny. The paper has concrete strengths: it uses publicly available pretrained models and datasets, it reports computational complexity for the frequency-domain operation, it compares across three imaging modalities, and it makes source code publicly available. However, the current evaluation protocol does not support the headline comparison, and the method definition in Eq. (2) leaves an essential implementation detail unspecified. The significance of the contribution is therefore conditional on correcting the evaluation protocol and clarifying the forward-pass definition.

major comments (3)

[Section 4, Fig. 2; Section 3.4] The central comparison in Fig. 2 is based on the best-performing dropout configuration for each method, selected using UCE computed on the same test cohorts used for evaluation. With 6 dropout rates and 3 placements, this is a minimum over 18 noisy per-instance UCE estimates per method, and minima are systematically biased downward by winner's-curse effects even when the methods are equivalent. The reported test cohort sizes (n=36, 37, 40) make this bias substantial. The paper describes no validation split, no error bars on the Fig. 2 UCE values, and no correction for multiple comparisons. Because Fig. 3 shows strong dependence on rate and placement, the selection step is exactly where the bias enters. The authors should either report the full per-configuration results with confidence intervals and significance tests, or use a held-out selection protocol in which configuration choice is made on data not used for the reported comparison.
[Section 2.2, Eq. (2)] Eq. (2) defines the frequency-diluted forward pass as C_F_theta(X) = sigma(F^{-1}(F[X * W] ⊙ D) + b), where D is a binary dropout mask. For a real-valued feature map, applying an arbitrary real binary mask to the full complex Fourier transform and then taking the inverse FFT produces a complex output unless the mask preserves Hermitian symmetry (D(k) = conj(D(-k))). The paper never states that D has this symmetry, and it does not specify whether the implementation uses a full FFT with a symmetric mask or an rfft/irfft pair. As written, the forward pass could produce complex feature maps, which would break the CNN. This is a load-bearing implementation detail for the method, and it must be stated explicitly in the paper.
[Table 1] The DSC divergence values in Table 1 are difficult to interpret. For liver tumors, the baseline DSC is 0.606, yet the reported 'Divergence from baseline DSC' values are -14.53 and -11.72 for Signal Dropout and 1.23 and 3.54 for Frequency Dropout. If these are percentage-point differences they are inconsistent with DSC being in [0,1]; if they are relative percentage changes the units need to be stated. The asterisk annotation is also applied only to some entries. Since the paper claims 'stable segmentation performance' partly on the basis of this table, the metric definition and its units must be clarified, and ideally the underlying DSC values should be reported.

minor comments (4)

[Fig. 3 caption] The caption of Fig. 3 refers to 'Expected Calibration Error (ECE)', but Section 3.5 defines the metric as Expected Uncertainty Calibration Error (UCE). These are different quantities, and the notation should be made consistent throughout.
[Section 2.2] The paper should specify whether F and F^{-1} denote the 2D DFT and its inverse with a particular normalization convention, and should state whether the implementation uses rfft/irfft on real feature maps or a full FFT with a Hermitian-symmetric mask.
[Section 5] The phrase 'convergence' in the abstract and discussion is not operationalized; no convergence metric or analysis is defined in Section 3.5 or shown in the results. Consider removing or explicitly defining this claim.
[Section 5, Discussion] There is a minor wording issue: 'should be weighed to determines suitability' should be 'should be weighed to determine suitability'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's derivation is self-contained, and its self-citation is not load-bearing.

full rationale

The paper contains no equation-level circularity. Equation (1) and Equation (2) define two distinct forward passes—masking in the signal domain versus masking in the Fourier domain—and the reported UCE and DSC values are empirical measurements, not identities that follow from those definitions. The frequency-diluted forward pass is attributed to the authors' prior ISBI paper [10], but Eq. (2) fully specifies the operation and public code is referenced, so this self-citation is not load-bearing; the evaluation uses independent public datasets and pretrained models (Medical Segmentation Decathlon and the chest X-ray database with nnU-Net and MedSAM). The main non-circularity concerns are methodological: Section 4 and Fig. 2 compare "best dropout configurations" without stating that the best configuration was selected on held-out data, so the reported minima may reflect selection bias, and Eq. (2)'s real binary mask applied to a complex Fourier transform requires Hermitian symmetry (or rfft/irfft) to produce real feature maps. Both are correctness risks rather than circular reductions, because neither makes the reported outcome equal to the method's defining equation by construction. No specific circular step can therefore be quoted.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical entities. The central evaluation rests on several modeling assumptions: the MC-dropout-as-Bayesian approximation, inference-time dropout on pretrained models, and an unstated symmetry-preserving frequency mask. The main free parameters are the dropout rate and layer placement, which are selected on the test set.

free parameters (3)

dropout rate p = selected per task and method from {0.01, 0.02, 0.04, 0.08, 0.16, 0.32}
The paper reports the 'best-performing' dropout rate for each method and task, chosen by evaluating UCE on the same test set (Section 3.4, Fig. 2), which acts as test-set fitting.
dropout layer placement = encoder, decoder, or global
The paper compares three placements and reports the best per method (Fig. 3), another test-set selection.
number of MC repetitions R = 30 (and 5 for comparison)
R=30 is chosen for the main results; R=5 is also shown in Table 1. This is a hand-chosen modeling parameter, not fitted to data.

assumptions (4)

domain assumption MC-dropout approximates Bayesian inference for uncertainty estimation (Gal and Ghahramani [9]).
The paper treats the standard deviation of MC predictions as predictive uncertainty, relying on this approximation without discussion (Section 3.4).
domain assumption Applying dropout to a pretrained model, without retraining, yields meaningful uncertainty estimates.
The paper uses pretrained nnU-Net and MedSAM models and never trains with dropout; this is a strong assumption about the validity of MC dropout at inference time.
ad hoc to paper The frequency mask in Eq. (2) yields a real feature map after inverse FFT.
The paper does not state that D must preserve Hermitian symmetry; the method as written may produce complex outputs, so a symmetry-preserving construction is an unstated ad hoc assumption.
domain assumption UCE computed against full-model segmentation errors on a single test set is an unbiased estimator of calibration.
The evaluation treats test-set UCE as the target metric without a validation split, which is violated by the test-set hyperparameter selection.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Enhancing Uncertainty Estimation in Semantic Segmentation via Monte-Carlo Frequency Dropout." pith.science (2026). https://pith.science/paper/YHPHADDD

@misc{pith2026250111258,
  author       = {Pith},
  title        = {Pith review of: Enhancing Uncertainty Estimation in Semantic Segmentation via Monte-Carlo Frequency Dropout},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YHPHADDD}},
  note         = {Machine review of arXiv:2501.11258}
}

read the original abstract

Monte-Carlo (MC) Dropout provides a practical solution for estimating predictive distributions in deterministic neural networks. Traditional dropout, applied within the signal space, may fail to account for frequency-related noise common in medical imaging, leading to biased predictive estimates. A novel approach extends Dropout to the frequency domain, allowing stochastic attenuation of signal frequencies during inference. This creates diverse global textural variations in feature maps while preserving structural integrity -- a factor we hypothesize and empirically show is contributing to accurately estimating uncertainties in semantic segmentation. We evaluated traditional MC-Dropout and the MC-frequency Dropout in three segmentation tasks involving different imaging modalities: (i) prostate zones in biparametric MRI, (ii) liver tumors in contrast-enhanced CT, and (iii) lungs in chest X-ray scans. Our results show that MC-Frequency Dropout improves calibration, convergence, and semantic uncertainty, thereby improving prediction scrutiny, boundary delineation, and has the potential to enhance medical decision-making.

Figures

Figures reproduced from arXiv: 2501.11258 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 20 canonical work pages

[11]

Accurate and ro- bust deep learning-based segmentation of the prostate clinical target volume in ultrasound images,

Davood Karimi, Qi Zeng, Prateek Mathur, Apeksha Avinash, Sara Mahdavi, Ingrid Spadinger, Purang Abol- maesumi, and Septimiu E Salcudean, “Accurate and ro- bust deep learning-based segmentation of the prostate clinical target volume in ultrasound images,” Medical image analysis, vol. 57, pp. 186–196, 2019

work page 2019
[10]

Automatic brain tumor segmentation us- ing convolutional neural networks with test-time aug- mentation,

Guotai Wang, Wenqi Li, S ´ebastien Ourselin, and Tom Vercauteren, “Automatic brain tumor segmentation us- ing convolutional neural networks with test-time aug- mentation,” in Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries: 4th International Workshop, BrainLes 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, Septemb...

work page 2018
[1]

Enhancing Uncertainty Estimation in Semantic Segmentation via Monte-Carlo Frequency Dropout

INTRODUCTION Estimating prediction uncertainties in deterministic deep learning models often involves the strategic introduction of controlled artificial noise into the data [1]. This can occur either before [2, 3] or during [4, 5, 6] neural network pro- cessing, with subsequent measurement of variations in model performance to assess robustness. Techniqu...

work page Pith review arXiv 2025
[2]

BACKGROUND Without loss of generality, let Cθ : Rm×n → Rm′×n′ denote a convolution block within a CNN model MΘ, comprising a single convolutional layer, where θ ∈ Θ represents the set of trainable model parameters. The forward pass output ofCθ for a given input instanceX ∈ Rm×n is expressed as follows: Cθ(X) = σ(X ∗ W + b) where W k×k ∈ θ represents the c...

work page
[3]

The CT and MRI datasets were sourced from the Medical Segmentation Decathlon [12]

EMPIRICAL EV ALUATION We compared uncertainty estimates from Monte Carlo (MC) simulations with Signal and Frequency Dropout during in- ference to identify segmentation errors across three public semantic segmentation tasks involving MRI, CT, and X-ray modalities. The CT and MRI datasets were sourced from the Medical Segmentation Decathlon [12]. The source...

work page
[4]

RESULTS Across all segmentation tasks, the best-performing Frequency Dropout configuration converged to better-calibrated uncer- tainty estimates than Signal Dropout, with uncertainties more closely matching full model segmentation errors (Fig. 2). Similar performance was observed for liver tumor segmen- tation. Frequency Dropout achieved higher or compar...

work page
[5]

MC simulations generated uncertainty estimates to identify segmentation er- rors in tasks challenging state-of-the-art models

DISCUSSION AND CONCLUSION This paper explores two Monte Carlo (MC) Dropout ap- proaches for medical imaging segmentation: traditional (Sig- nal) Dropout, applied directly to feature maps, and Frequency Dropout, operating in the frequency domain. MC simulations generated uncertainty estimates to identify segmentation er- rors in tasks challenging state-of-...

work page
[6]

COMPLIANCE WITH ETHICAL STANDARDS This research study was conducted retrospectively using hu- man subject data made available in open access by [12, 13, 14]

work page

Show all 28 references

[7]

The authors have no relevant financial or non-financial interests to dis- close

ACKNOWLEDGMENTS No funding was received to conduct this study. The authors have no relevant financial or non-financial interests to dis- close
[8]

A review of uncertainty quantification in med- ical image analysis: probabilistic and non-probabilistic methods,

Ling Huang, Su Ruan, Yucheng Xing, and Mengling Feng, “A review of uncertainty quantification in med- ical image analysis: probabilistic and non-probabilistic methods,” Medical Image Analysis, p. 103223, 2024

2024
[9]

Aleatoric uncertainty estimation with test-time augmentation for medical image segmentation with convolutional neural networks,

Guotai Wang, Wenqi Li, Michael Aertsen, Jan Deprest, S´ebastien Ourselin, and Tom Vercauteren, “Aleatoric uncertainty estimation with test-time augmentation for medical image segmentation with convolutional neural networks,” Neurocomputing, vol. 338, pp. 34–45, 2019

2019
[12]

Exploring uncertainty measures in bayesian deep attentive neural networks for prostate zonal segmentation,

Yongkai Liu, Guang Yang, Melina Hosseiny, Afshin Azadikhah, Sohrab Afshari Mirak, Qi Miao, Steven S Raman, and Kyunghyun Sung, “Exploring uncertainty measures in bayesian deep attentive neural networks for prostate zonal segmentation,” Ieee Access, vol. 8, pp. 151817–151828, 2020

2020
[13]

Exploring uncertainty measures in deep net- works for multiple sclerosis lesion detection and seg- mentation,

Tanya Nair, Doina Precup, Douglas L Arnold, and Tal Arbel, “Exploring uncertainty measures in deep net- works for multiple sclerosis lesion detection and seg- mentation,” Medical image analysis , vol. 59, pp. 101557, 2020

2020
[14]

Regularization of neural networks using dropconnect,

Li Wan, Matthew Zeiler, Sixin Zhang, Yann Le Cun, and Rob Fergus, “Regularization of neural networks using dropconnect,” in International conference on machine learning. PMLR, 2013, pp. 1058–1066

2013
[15]

Dropout: a simple way to prevent neural networks from overfit- ting,

Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov, “Dropout: a simple way to prevent neural networks from overfit- ting,” The journal of machine learning research , vol. 15, no. 1, pp. 1929–1958, 2014

1929
[16]

Dropout as a Bayesian approximation: Representing model uncer- tainty in deep learning,

Yarin Gal and Zoubin Ghahramani, “Dropout as a Bayesian approximation: Representing model uncer- tainty in deep learning,” in international conference on machine learning. PMLR, 2016, pp. 1050–1059

2016
[17]

Monte-carlo frequency dropout for predictive uncertainty estimation in deep learning,

Tal Zeevi, Rajesh Venkataraman, Lawrence H Staib, and John A Onofrey, “Monte-carlo frequency dropout for predictive uncertainty estimation in deep learning,” in 2024 IEEE International Symposium on Biomedical Imaging (ISBI). IEEE, 2024, pp. 1–5

2024
[18]

Regularization of deep neural networks with spectral dropout,

Salman H Khan, Munawar Hayat, and Fatih Porikli, “Regularization of deep neural networks with spectral dropout,” Neural Networks, vol. 110, pp. 82–90, 2019

2019
[19]

The medical seg- mentation decathlon,

Michela Antonelli, Annika Reinke, Spyridon Bakas, Keyvan Farahani, Annette Kopp-Schneider, Bennett A Landman, Geert Litjens, Bjoern Menze, Olaf Ron- neberger, Ronald M Summers, et al., “The medical seg- mentation decathlon,” Nature communications, vol. 13, no. 1, pp. 4128, 2022

2022
[20]

Lung segmentation in chest radiographs using anatomical atlases with nonrigid reg- istration,

Sema Candemir, Stefan Jaeger, Kannappan Palaniap- pan, Jonathan P Musco, Rahul K Singh, Zhiyun Xue, Alexandros Karargyris, Sameer Antani, George Thoma, and Clement J McDonald, “Lung segmentation in chest radiographs using anatomical atlases with nonrigid reg- istration,” IEEE ...

2013
[21]

Automatic tuberculosis screening using chest radiographs,

Stefan Jaeger, Alexandros Karargyris, Sema Candemir, Les Folio, Jenifer Siegelman, Fiona Callaghan, Zhiyun Xue, Kannappan Palaniappan, Rahul K Singh, Sameer Antani, et al., “Automatic tuberculosis screening using chest radiographs,” IEEE transactions on medical imag- ing, vol....

2013
[22]

nnu-net: a self- configuring method for deep learning-based biomedical image segmentation,

Fabian Isensee, Paul F Jaeger, Simon AA Kohl, Jens Petersen, and Klaus H Maier-Hein, “nnu-net: a self- configuring method for deep learning-based biomedical image segmentation,” Nature methods, vol. 18, no. 2, pp. 203–211, 2021

2021
[23]

Segment anything in medical images,

Jun Ma, Yuting He, Feifei Li, Lin Han, Chenyu You, and Bo Wang, “Segment anything in medical images,” Nature Communications, vol. 15, no. 1, pp. 654, 2024

2024
[24]

Segment anything,

Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alexander C Berg, Wan-Yen Lo, et al., “Segment anything,” in Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, 2023, pp. 4015–4026

2023
[25]

Evaluating bayesian deep learning methods for semantic segmentation,

Jishnu Mukhoti and Yarin Gal, “Evaluating bayesian deep learning methods for semantic segmentation,” arXiv preprint arXiv:1811.12709, 2018

2018 arXiv
[26]

Dropconnect is effective in modeling uncertainty of bayesian deep networks,

Aryan Mobiny, Pengyu Yuan, Supratik K Moulik, Naveen Garg, Carol C Wu, and Hien Van Nguyen, “Dropconnect is effective in modeling uncertainty of bayesian deep networks,” Scientific reports, vol. 11, no. 1, pp. 5458, 2021

2021
[27]

Well-calibrated model uncertainty with temperature scaling for dropout vari- ational inference,

Max-Heinrich Laves, Sontje Ihler, Karl-Philipp Kort- mann, and Tobias Ortmaier, “Well-calibrated model uncertainty with temperature scaling for dropout vari- ational inference,” arXiv preprint arXiv:1909.13550 , 2019

1909 arXiv
[28]

On calibration of modern neural networks,

Chuan Guo, Geoff Pleiss, Yu Sun, and Kilian Q Wein- berger, “On calibration of modern neural networks,” in International conference on machine learning . PMLR, 2017, pp. 1321–1330

2017

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [11]

Accurate and ro- bust deep learning-based segmentation of the prostate clinical target volume in ultrasound images,

Davood Karimi, Qi Zeng, Prateek Mathur, Apeksha Avinash, Sara Mahdavi, Ingrid Spadinger, Purang Abol- maesumi, and Septimiu E Salcudean, “Accurate and ro- bust deep learning-based segmentation of the prostate clinical target volume in ultrasound images,” Medical image analysis, vol. 57, pp. 186–196, 2019

work page 2019

[2] [10]

Automatic brain tumor segmentation us- ing convolutional neural networks with test-time aug- mentation,

Guotai Wang, Wenqi Li, S ´ebastien Ourselin, and Tom Vercauteren, “Automatic brain tumor segmentation us- ing convolutional neural networks with test-time aug- mentation,” in Brainlesion: Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries: 4th International Workshop, BrainLes 2018, Held in Conjunction with MICCAI 2018, Granada, Spain, Septemb...

work page 2018

[3] [1]

Enhancing Uncertainty Estimation in Semantic Segmentation via Monte-Carlo Frequency Dropout

INTRODUCTION Estimating prediction uncertainties in deterministic deep learning models often involves the strategic introduction of controlled artificial noise into the data [1]. This can occur either before [2, 3] or during [4, 5, 6] neural network pro- cessing, with subsequent measurement of variations in model performance to assess robustness. Techniqu...

work page Pith review arXiv 2025

[4] [2]

BACKGROUND Without loss of generality, let Cθ : Rm×n → Rm′×n′ denote a convolution block within a CNN model MΘ, comprising a single convolutional layer, where θ ∈ Θ represents the set of trainable model parameters. The forward pass output ofCθ for a given input instanceX ∈ Rm×n is expressed as follows: Cθ(X) = σ(X ∗ W + b) where W k×k ∈ θ represents the c...

work page

[5] [3]

The CT and MRI datasets were sourced from the Medical Segmentation Decathlon [12]

EMPIRICAL EV ALUATION We compared uncertainty estimates from Monte Carlo (MC) simulations with Signal and Frequency Dropout during in- ference to identify segmentation errors across three public semantic segmentation tasks involving MRI, CT, and X-ray modalities. The CT and MRI datasets were sourced from the Medical Segmentation Decathlon [12]. The source...

work page

[6] [4]

RESULTS Across all segmentation tasks, the best-performing Frequency Dropout configuration converged to better-calibrated uncer- tainty estimates than Signal Dropout, with uncertainties more closely matching full model segmentation errors (Fig. 2). Similar performance was observed for liver tumor segmen- tation. Frequency Dropout achieved higher or compar...

work page

[7] [5]

MC simulations generated uncertainty estimates to identify segmentation er- rors in tasks challenging state-of-the-art models

DISCUSSION AND CONCLUSION This paper explores two Monte Carlo (MC) Dropout ap- proaches for medical imaging segmentation: traditional (Sig- nal) Dropout, applied directly to feature maps, and Frequency Dropout, operating in the frequency domain. MC simulations generated uncertainty estimates to identify segmentation er- rors in tasks challenging state-of-...

work page

[8] [6]

COMPLIANCE WITH ETHICAL STANDARDS This research study was conducted retrospectively using hu- man subject data made available in open access by [12, 13, 14]

work page