Distance-based detection of out-of-distribution silent failures for Covid-19 lung lesion segmentation

González, Camila; Gotkowski, Karol; Fuchs, Moritz; Bucher, Andreas; Dadras, Armin; Fischbach, Ricarda; Kaltenborn, Isabel; Mukhopadhyay, Anirban

doi:10.1016/j.media.2022.102596

Cited by 16 publications

(4 citation statements)

References 27 publications

(38 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This potential advantage of the U LG measure should be considered within the context of the image analysis task. The work of (González et al 2022) similarly demonstrated the utility of an uncertainty measure using artificial image degradations, however, this study specifically targeted OOD uncertainty (a form of epistemic uncertainty). In contrast, we designed a method to target uncertainty more generally.…”

Section: Discussionmentioning

confidence: 99%

“…Several studies have begun implementing UQ methods for medical image analysis tasks. Among these medical image studies, reporting the maximum softmax probability as a baseline is common (DeVries and Taylor 2018, Berger et al 2021, González et al 2022. The entropy of predicted probabilities has been used for UQ in structure delineation tasks for PET, CT, and MRI data (Mehrtash et al 2020, Diao et al 2022.…”

Section: Introduction 1overviewmentioning

confidence: 99%

See 1 more Smart Citation

Uncertainty quantification via localized gradients for deep learning-based medical image assessments

Schott,

Pinchuk,

Santoro-Fernandes

et al. 2024

Phys. Med. Biol.

View full text Add to dashboard Cite

Objective. Deep learning models that aid in medical image assessment tasks must be both accurate and reliable to be deployed within clinical settings. While deep learning models have been shown to be highly accurate across a variety of tasks, measures that indicate the reliability of these models are less established. Increasingly, uncertainty quantification (UQ) methods are being introduced to inform users on the reliability of model outputs. However, most existing methods cannot be augmented to previously validated models because they are not post hoc, and they change a model’s output. In this work, we overcome these limitations by introducing a novel post hoc UQ method, termed Local Gradients UQ, and demonstrate its utility for deep learning-based metastatic disease delineation. Approach. This method leverages a trained model’s localized gradient space to assess sensitivities to trained model parameters. We compared the Local Gradients UQ method to non-gradient measures defined using model probability outputs. The performance of each uncertainty measure was assessed in four clinically relevant experiments: (1) response to artificially degraded image quality, (2) comparison between matched high- and low-quality clinical images, (3) false positive (FP) filtering, and (4) correspondence with physician-rated disease likelihood. Main results. (1) Response to artificially degraded image quality was enhanced by the Local Gradients UQ method, where the median percent difference between matching lesions in non-degraded and most degraded images was consistently higher for the Local Gradients uncertainty measure than the non-gradient uncertainty measures (e.g., 62.35% vs. 2.16% for additive Gaussian noise). (2) The Local Gradients UQ measure responded better to high- and low-quality clinical images (p<0.05 vs p>0.1 for both non-gradient uncertainty measures). (3) FP filtering performance was enhanced by the Local Gradients UQ method when compared to the non-gradient methods, increasing the area under the receiver operating characteristic curve (ROC AUC) by 20.1% and decreasing the false positive rate by 26%. (4) The Local Gradients UQ method also showed more favorable correspondence with physician-rated likelihood for malignant lesions by increasing ROC AUC for correspondence with physician-rated disease likelihood by 16.2%. Significance. In summary, this work introduces and validates a novel gradient-based UQ method for deep learning-based medical image assessments to enhance user trust when using deployed clinical models.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introduction 1overviewmentioning

confidence: 99%

Uncertainty quantification via localized gradients for deep learning-based medical image assessments

Schott,

Pinchuk,

Santoro-Fernandes

et al. 2024

Phys. Med. Biol.

View full text Add to dashboard Cite

show abstract

“…Training on auxiliary negative data may give rise to undesirable bias and over-optimistic evaluation. Moreover, appropriate negative data may be unavailable in some application areas, such as medical diagnostics [ 24 ] or remote sensing [ 16 , 25 ]. Our experiments suggest that synthetic negatives may help in such cases.…”

Section: Introductionmentioning

confidence: 99%

Dense Out-of-Distribution Detection by Robust Learning on Synthetic Negative Data

Grcić,

Bevandić,

Kalafatić

et al. 2024

Sensors

View full text Add to dashboard Cite

Standard machine learning is unable to accommodate inputs which do not belong to the training distribution. The resulting models often give rise to confident incorrect predictions which may lead to devastating consequences. This problem is especially demanding in the context of dense prediction since input images may be only partially anomalous. Previous work has addressed dense out-of-distribution detection by discriminative training with respect to off-the-shelf negative datasets. However, real negative data may lead to over-optimistic evaluation due to possible overlap with test anomalies. To this end, we extend this approach by generating synthetic negative patches along the border of the inlier manifold. We leverage a jointly trained normalizing flow due to a coverage-oriented learning objective and the capability to generate samples at different resolutions. We detect anomalies according to a principled information-theoretic criterion which can be consistently applied through training and inference. The resulting models set the new state of the art on benchmarks for out-of-distribution detection in road-driving scenes and remote sensing imagery despite minimal computational overhead.

show abstract

“…This problem, commonly known as out-of-distribution (OoD) detection [ 1 ], aims to alleviate the risks of evaluating OoD samples, as performances on these are known to be erratic and typically produce wrong answers with high confidences, whereby making them potentially dangerous. As machine learning has become increasingly prevalent in mission-critical systems, the problem of OoD detection has gathered significant attention both in general computer vision research [ 1 ], and in applied medical imaging systems [ 2 – 8 ].…”

Section: Introductionmentioning

confidence: 99%

Unsupervised out-of-distribution detection for safer robotically guided retinal microsurgery

et al. 2023

View full text Add to dashboard Cite

Purpose A fundamental problem in designing safe machine learning systems is identifying when samples presented to a deployed model differ from those observed at training time. Detecting so-called out-of-distribution (OoD) samples is crucial in safety-critical applications such as robotically guided retinal microsurgery, where distances between the instrument and the retina are derived from sequences of 1D images that are acquired by an instrument-integrated optical coherence tomography (iiOCT) probe. Methods This work investigates the feasibility of using an OoD detector to identify when images from the iiOCT probe are inappropriate for subsequent machine learning-based distance estimation. We show how a simple OoD detector based on the Mahalanobis distance can successfully reject corrupted samples coming from real-world ex vivo porcine eyes. Results Our results demonstrate that the proposed approach can successfully detect OoD samples and help maintain the performance of the downstream task within reasonable levels. MahaAD outperformed a supervised approach trained on the same kind of corruptions and achieved the best performance in detecting OoD cases from a collection of iiOCT samples with real-world corruptions. Conclusion The results indicate that detecting corrupted iiOCT data through OoD detection is feasible and does not need prior knowledge of possible corruptions. Consequently, MahaAD could aid in ensuring patient safety during robotically guided microsurgery by preventing deployed prediction models from estimating distances that put the patient at risk.

show abstract

Distance-based detection of out-of-distribution silent failures for Covid-19 lung lesion segmentation

Cited by 16 publications

References 27 publications

Uncertainty quantification via localized gradients for deep learning-based medical image assessments

Uncertainty quantification via localized gradients for deep learning-based medical image assessments

Dense Out-of-Distribution Detection by Robust Learning on Synthetic Negative Data

Unsupervised out-of-distribution detection for safer robotically guided retinal microsurgery

Contact Info

Product

Resources

About