Pith. sign in

REVIEW 13 cited by

Evaluating Bayesian Deep Learning Methods for Semantic Segmentation

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1811.12709 v2 pith:SLZEUUML submitted 2018-11-30 cs.CV

Jishnu Mukhoti , Yarin Gal This is my paper · ORCID
classification cs.CV
keywords semanticsegmentationdeepmetricsbayesianevaluatelearninguncertainty
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Deep learning has been revolutionary for computer vision and semantic segmentation in particular, with Bayesian Deep Learning (BDL) used to obtain uncertainty maps from deep models when predicting semantic classes. This information is critical when using semantic segmentation for autonomous driving for example. Standard semantic segmentation systems have well-established evaluation metrics. However, with BDL's rising popularity in computer vision we require new metrics to evaluate whether a BDL method produces better uncertainty estimates than another method. In this work we propose three such metrics to evaluate BDL models designed specifically for the task of semantic segmentation. We modify DeepLab-v3+, one of the state-of-the-art deep neural networks, and create its Bayesian counterpart using MC dropout and Concrete dropout as inference techniques. We then compare and test these two inference techniques on the well-known Cityscapes dataset using our suggested metrics. Our results provide new benchmarks for researchers to compare and evaluate their improved uncertainty quantification in pursuit of safer semantic segmentation.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 13 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Enhancing Uncertainty Estimation and Interpretability via Bayesian Non-negative Decision Layer

    cs.LG 2025-05 conditional novelty 6.0 of 10

    A Bayesian non-negative decision layer with gamma priors and Weibull variational inference improves uncertainty estimation and interpretability for image classifiers.

  2. Rate-In: Information-Driven Adaptive Dropout Rates for Improved Inference-Time Uncertainty Estimation

    cs.LG 2024-12 conditional novelty 6.0 of 10

    Rate-In adapts dropout rates per layer and per input at inference by measuring mutual-information loss, improving calibration and sharpening uncertainty estimates in medical imaging tests.

  3. Segment-Level Road Obstacle Detection Using Visual Foundation Model Priors and Likelihood Ratios

    cs.CV 2024-12 conditional novelty 6.0 of 10

    A segment-level road obstacle detector using SAM features and a likelihood ratio of free-space versus obstacle distributions achieves state-of-the-art component-level scores on RoadObstacle and LostAndFound.

  4. UCBound-Net: Uncertainty-Guided Boundary-Aware Continual Learning for Domain-Incremental Ultrasound Segmentation

    cs.CV 2026-08 conditional novelty 5.0 of 10

    UCBound-Net uses MC Dropout uncertainty maps to focus distillation, calibration, and replay on boundary regions, reducing catastrophic forgetting in domain-incremental ultrasound segmentation.

  5. Spatially-Aware Evaluation of Segmentation Uncertainty

    cs.CV 2025-06 conditional novelty 5.0 of 10

    New spatially aware metrics (BUC, BA-ECE, SPACE) evaluate segmentation uncertainty using boundary distance and local smoothing, and outperform voxel-wise metrics at distinguishing clean from noisy uncertainty maps on ...

  6. Rethinking Semi-supervised Segmentation Beyond Accuracy: Reliability and Robustness

    cs.CV 2025-06 conditional novelty 5.0 of 10

    A new harmonic-mean metric, RSS, combines mIoU, calibration error, and two uncertainty-quality measures, and is used to show that SSL segmentation models like UniMatchV2 sacrifice reliability for accuracy.

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

    cs.CV 2025-01 conditional novelty 5.0 of 10

    Monte-Carlo Frequency Dropout, which randomly removes frequency components in feature maps, yields better-calibrated uncertainty estimates than standard signal dropout for medical image segmentation.

  8. Uncertainty in Real-Time Semantic Segmentation on Embedded Systems

    cs.CV 2022-12 unverdicted novelty 5.0 of 10

    Combines pre-trained features, Bayesian regression, and moment propagation to enable real-time epistemic uncertainty for semantic segmentation on embedded systems while preserving accuracy.

  9. Instance-Level Post Hoc Uncertainty Quantification in Object Detection

    cs.CV 2026-06 unverdicted novelty 4.0 of 10

    Proposes MC-GLM to deliver instance-level post-hoc uncertainty for object detectors via Laplace approximation plus constant-cost Monte Carlo sampling.

  10. A Critical Synthesis of Uncertainty Quantification and Foundation Models in Monocular Depth Estimation

    cs.CV 2025-01 conditional novelty 4.0 of 10

    Fine-tuning DepthAnythingV2 with a Gaussian negative log-likelihood loss yields the most reliable pixel-wise uncertainty estimates on indoor, street, and object scenes, but it fails on aerial large-depth data.

  11. Safety Monitoring of Machine Learning Perception Functions: a Survey

    cs.LG 2024-12 accept novelty 4.0 of 10

    A survey that organizes research on runtime safety monitors for ML perception into threat identification, requirements, detection, reaction, and evaluation, and lists open challenges.

  12. Uncertainty-Aware Artificial Intelligence for Gear Fault Diagnosis in Motor Drives

    eess.SY 2024-12 conditional novelty 4.0 of 10

    A Bayesian neural network with Bayes-by-backprop is used for gear fault diagnosis, and its uncertainty estimates increase for noisy data and unseen fault classes.

  13. Distributional Uncertainty for Out-of-Distribution Detection

    cs.CV 2025-07 reject novelty 3.0 of 10

    A new loss and uncertainty map built from a Beta posterior network aim to improve out-of-distribution pixel detection for semantic segmentation, but the reported gains are inconsistent and the novelty is limited.

Pith tools