Making the Flow Glow -- Robot Perception under Severe Lighting Conditions using Normalizing Flow Gradients

Rudolph Triebel; Simon Kristoffersson Lind; Volker Kr\"uger

REVIEW 3 major objections 5 minor 29 references

Making the Flow Glow -- Robot Perception under Severe Lighting Conditions using Normalizing Flow Gradients

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read This paper claims that the absolute pixel-wise gradients of a normalizing flow carry local out-of-distribution information, and that optimizing camera parameters against this local score lifts object-detection success by 60% over the…

desk verdict A plausible new trick for local OOD scoring in robot perception, but the experiments don't yet isolate why it works. read the letter →

arxiv 2412.07565 v1 pith:RD56KFAL submitted 2024-12-10 cs.CV

Simon Kristoffersson Lind , Rudolph Triebel , Volker Krüger This is my paper · ORCID

classification cs.CV

keywords out-of-distributiondetectionnormalizingflowpixel-wisegradientscameraparameteroptimizationroboticperceptionobjectseverelightingregionofinterest

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 tries to establish that the absolute pixel-wise gradients of a normalizing flow, computed by backpropagating through the flow and a feature extractor, carry local out-of-distribution information that a global likelihood score cannot provide. If true, a robot can optimize camera parameters so that the detector's proposed regions of interest become visible, rather than trading off visibility across the whole image. The authors support this with a conventional OOD detection benchmark, a correlation study on COCO bounding boxes, and a tabletop picking experiment under severe lighting where their method achieves 60% more correct detections than the previous likelihood-based approach.

What carries the argument

The central object is the absolute NF gradient image, defined by backpropagating through the composed map $x = F(I)$ and then through the normalizing flow $T$ to obtain $|\delta p_{train}/\delta i_{yx}|$ for each pixel and taking absolute values. The rationale is that in-distribution inputs lie near a local maximum of the learned density, where gradients are small. The optimization objective averages this gradient over the union of detector-proposed regions of interest, so the camera parameters are tuned to make exactly those regions look in-distribution.

What would settle it

Take a collection of images where the objects are out-of-distribution for the flow but a cluttered background is in-distribution, and measure the absolute NF gradient inside correctly detected bounding boxes versus inside false positives; a finding that correct detections have consistently higher absolute gradients than false positives would contradict the core assumption.

Watch

Extended reading notes

Core claim

The central claim is that the map $|\nabla p_{train}(F(I))|$, the absolute value of the gradient of the normalizing-flow-estimated density with respect to input pixels, acts as a pixel-level OOD score: in-distribution image regions tend to have small absolute gradient values and out-of-distribution regions large ones. Using this map, the robot minimizes the average absolute gradient inside the regions proposed by an object detector, with a global-average fallback when no region is proposed. In a dark room with a strong upward light, this optimization yields 114 correct object detections across 83 trials, compared to 71 for the likelihood-based method of [3] and 45 for default camera parameters.

Load-bearing premise

The whole method rests on the observed correlation that image regions matching the training distribution have smaller absolute normalizing-flow gradients than out-of-distribution regions, so if that ordering ever reversed, optimizing the gradient would not improve detection.

Editorial extensions

If this is right

Robots can adapt camera exposure, gain, and related parameters to make proposed detection regions legible even when no global setting makes the whole scene visible.
The same gradient map can serve as a fallback OOD detector when no region proposal exists, since its global average retains OOD detection ability close to that of the likelihood itself.
Object detectors, including YOLOv4 and Faster-RCNN, tend to predict correct bounding boxes in regions with lower absolute NF gradients, linking local OOD score to downstream task success.
Per-object optimization, rather than joint optimization over all proposed regions, is identified by the authors as a likely further improvement.

Reading between the lines

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

A likely consequence the authors do not explore is that replacing the evolutionary camera-parameter search with a learned or analytic mapping from gradient statistics to parameters could remove the slow camera-sampling dependency they cite as a limitation.
The stated rationale that in-distribution points lie near maxima is a heuristic; the gradient map could still be a useful OOD signal even if that geometric picture fails, because it measures the local sensitivity of the density rather than its level.
Since the gradient is computed through a pretrained feature extractor, the method inherits the extractor's notion of distribution, so swapping or retraining the extractor offers a testable route to transferring the approach to new domains or architectures.

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: the core idea—using absolute normalizing flow gradients as a pixel-level OOD signal and optimizing camera parameters over detector ROIs—is genuinely new and worth taking seriously. The paper does not, however, establish the mechanism it claims. The 60% improvement is real in the narrow sense that 114 vs 71 detections, but the comparison bundles two changes at once, and the load-bearing assumption that low gradient marks in-distribution regions is only heuristically motivated and validated indirectly.

What the paper does well: it identifies a real failure mode in prior work (global likelihood optimization can't fix local lighting) and proposes a sensible fix. The gradient map is a natural way to localize OOD signal. The robot experiment, 83 trials with an actual arm and camera, is more than most papers of this type do. The authors are also transparent about runtime limits and the possibility that per-object optimization could improve results.

Where the evidence is soft: first, the premise itself. The intuition that in-distribution points lie near maxima of the learned density is not a consequence of density estimation; flat low-density regions have small gradients too. The averaged-gradient OOD tables (I-II) test whole-image discrimination, not localization. Table III shows correct boxes having ~20% lower mean gradient, but no variances, sample counts, or significance tests, and the authors' statement that standard deviations are 'meaningless' for non-Gaussian data is not a reason to skip interval estimates. Second, the robot comparison in Table IV is confounded: their method differs from [3] in both the objective (gradient vs likelihood) and the domain (ROI vs whole image), so the 60% could come from ROI focusing, evolutionary search noise, or CNN Jacobian artifacts rather than from genuine OOD semantics. Third, apple and mouse are excluded post-hoc; that is defensible but should be stated as an exclusion criterion, not an afterthought. Fourth, promised code and data are not yet available.

None of this is fatal to the idea—I'd bet the gradient signal carries some real local OOD information—but the paper as written supports a conditional acceptance, not a settled result.

Who it's for: robotics people working on active perception, camera self-calibration, and OOD-detection-for-adaptation. It would make a fine reading-group discussion piece. I would not cite the 60% number in my own writing yet, but I'd cite the method if code appears. It deserves a serious peer review; the right outcome would be a revise that adds ablations (e.g., gradient-global vs likelihood-ROI), bootstrap intervals, and a direct region-level test of the gradient assumption.

Referee Report

3 major / 5 minor

Summary. The paper proposes using the absolute pixel-wise gradient of a normalizing flow (NF) density, computed on features of a pretrained CNN, as a local out-of-distribution (OOD) score. This score is used to optimize camera parameters (backlight compensation, brightness, contrast, exposure, gain, saturation, sharpness) over the region of interest (ROI) proposed by an object detector, with a global-average fallback when no ROI is available. The authors evaluate the average absolute NF gradient as a global OOD score on CIFAR-10/SVHN/CelebA (Tables I-II), compare mean gradients inside correct vs. incorrect COCO bounding boxes for YOLOv4 and Faster-RCNN (Table III), and report a real robot tabletop picking experiment under severe lighting in which the proposed method yields 114 correct detections versus 71 for the likelihood-based baseline [3] (Table IV).

Significance. If the central premise holds, the work is a practical and relevant contribution: it extends NF-based OOD detection to pixel-level scoring and demonstrates a robotic camera-adaptation system that improves object detection in challenging lighting. The experiments are conducted on a real robot with a direct comparison to prior work, and the planned public release of code and data is commendable. The paper's main weakness is that the load-bearing assumption — small absolute NF gradients mark in-distribution image regions — is supported only by indirect, aggregate evidence without uncertainty quantification, and the robot experiment confounds the gradient score with ROI-based optimization. The novelty of the idea and the system-level demonstration justify a major revision rather than rejection, provided the mechanism is more rigorously validated.

major comments (3)

[Section IV, Eq. (2) and Section V-C, Table III] The central claim that small absolute NF gradient values indicate in-distribution regions is supported only by whole-image average OOD scores (Tables I-II) and by mean gradient differences inside correct vs. incorrect bounding boxes (Table III). Table III reports no variance, sample size, or significance test, and the statement that standard deviations are 'meaningless' for non-Gaussian data is not a valid substitute for interval estimates or nonparametric tests. As written, this evidence does not establish that the gradient score localizes OOD regions at the pixel or box level, which is the mechanism the optimization in Eq. (2) relies on. Please provide a region-level evaluation (e.g., patch-level ground truth), or at minimum report per-box distributions with bootstrap confidence intervals and a significance test (e.g., Mann-Whitney U) for Table III. Also, the heuristic that in-distribution points lie near a maximum of p_train is not theoretically guaranteed; flat low-density plateaus can also produce small gradients. A controlled experiment with a known density or a synthetic OOD region would substantially strengthen this core premise.
[Section V-D, Table IV] The robot experiment confounds the proposed gradient-based local score with the choice of optimization region. The baseline [3] optimizes the global likelihood over the entire image, while the proposed method optimizes the average absolute gradient over the detector's ROI (with a global fallback). The observed 60% improvement (114 vs. 71) could therefore stem from restricting optimization to the ROI, from the evolutionary search's stochasticity, or from the gradient objective's sensitivity to high-frequency image artifacts, rather than from the OOD semantics of the absolute NF gradient. To attribute the improvement to the proposed score, please add an ablation that optimizes the original likelihood over the same ROI (and, ideally, the average absolute gradient over the whole image). Without such an ablation, the central claim of superiority of the gradient-based local score is not established.
[Section V-A and V-B/D] The NF training and feature extraction details are incomplete for the COCO experiments (Table III) and the robot experiment (Table IV). The paper specifies only 10 affine coupling layers, 200 epochs, Adam with learning rate 1e-4, but not the input feature dimensionality, which layer of the pretrained CNN is used as F, whether features are preprocessed, or how the NF is trained for the COCO feature space. For the robot experiment, it is also unclear whether the NF was trained on COCO images, on the feature distribution of YOLOv4's convolutional layers, or on something else. These details are essential for reproducibility and for assessing whether Table III's correlation is meaningful. Please add a dedicated experimental setup subsection covering all models and datasets used for each experiment.

minor comments (5)

[Section V-B] The OOD experiments in Tables I-II evaluate the average absolute gradient over entire images; this is a global score and does not directly test the spatial localization property that the paper motivates. Please state explicitly that this experiment is only a sanity check for the global fallback, and that the localization evidence is provided (or will be provided) by the proposed region-level analysis.
[Section V-C] The sentence 'standard deviations appear meaningless' is not an adequate justification for omitting all dispersion information. Consider reporting interquartile ranges or bootstrap confidence intervals, which do not assume Gaussianity.
[Fig. 2] The caption describes 'absolute gradient images' but does not indicate which input images correspond to in-distribution or OOD examples. Please add annotations or labels to make the claimed visual effect interpretable to the reader.
[Section IV] The phrase 'in-distribution data will likely lie close to a maximum of p_train(x)' is presented as intuition. Please either soften the wording ('we hypothesize') or provide a more formal justification, as this is the crux of the proposed method.
[Abstract and Section V-D] The '60% higher success rate' refers specifically to the comparison against the NF baseline [3], not against all 'previous methods' in a broad sense. Consider clarifying this in the abstract to avoid overgeneralization.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the NF-gradient score is evaluated on held-out data and the 60% robot gain is an empirical comparison, not an equation-level reduction.

full rationale

The paper's central derivation is self-contained: Eq. (2) defines a new local score by backpropagating through a fixed NF and pretrained feature extractor; no parameter is fitted to the detection outcome. The OOD benchmark (Tables I-II) evaluates this score on held-out CIFAR/SVHN/CelebA data, and Table III is a correlation check, not a fitted prediction. The robot experiment (Table IV) compares the gradient objective against the global-likelihood baseline [3] using identical camera parameters and evolutionary hyperparameters, which makes the comparison fair rather than circular. The heuristic assumption that in-distribution points sit near density maxima (Section IV) is an unproven empirical premise and a correctness risk, but it is not equivalent by construction to the claimed improvement. The only self-citation, [3], supplies the baseline and optimization setup; it does not define the gradient score or forbid alternatives. Self-citation is therefore not load-bearing. No circular step can be exhibited from the paper's equations.

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

The method rests on two empirical assumptions: the gradient-OOD correlation and the benefit of ROI-local optimization. Standard NF mathematics and pretrained feature extractors are drawn from prior literature. The optimization and training hyperparameters are chosen by hand but not fitted to the target result.

free parameters (3)

Evolutionary optimization hyperparameters = population=50, mutation rate=20%, 200 iterations
Chosen identically to prior work [3]; not fitted to this paper's target result, but they affect optimization behavior and are not derived.
NF training hyperparameters = 200 epochs, Adam lr=1e-4
Hand-selected; not fitted to the target result.
NF architecture capacity
10 affine coupling layers with two shared ReLU linear layers; hidden dimensions not given, so capacity is unspecified.

assumptions (4)

domain assumption Absolute NF gradient values are smaller for in-distribution regions than for OOD regions.
Introduced in Section IV as an empirical observation from initial testing, with a heuristic justification. This is the load-bearing premise for using gradients as a local OOD score.
domain assumption Minimizing average absolute gradient over detector ROIs improves detection.
The basis for the robot experiment; supported only by the 83-trial result, not by a proof or controlled baseline.
domain assumption Pretrained CNN features provide a suitable space for NF-based OOD detection.
Adopted from [17] and [3]; the method inherits this assumption.
standard math Normalizing flow likelihood formula (Eq. 1) holds with the learned transformation.
Standard result from the NF literature [18].

how reviews work

0 comments

Cite this review

Pith. "Pith review of Making the Flow Glow -- Robot Perception under Severe Lighting Conditions using Normalizing Flow Gradients." pith.science (2026). https://pith.science/paper/RD56KFAL

@misc{pith2026241207565,
  author       = {Pith},
  title        = {Pith review of: Making the Flow Glow -- Robot Perception under Severe Lighting Conditions using Normalizing Flow Gradients},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RD56KFAL}},
  note         = {Machine review of arXiv:2412.07565}
}

read the original abstract

Modern robotic perception is highly dependent on neural networks. It is well known that neural network-based perception can be unreliable in real-world deployment, especially in difficult imaging conditions. Out-of-distribution detection is commonly proposed as a solution for ensuring reliability in real-world deployment. Previous work has shown that normalizing flow models can be used for out-of-distribution detection to improve reliability of robotic perception tasks. Specifically, camera parameters can be optimized with respect to the likelihood output from a normalizing flow, which allows a perception system to adapt to difficult vision scenarios. With this work we propose to use the absolute gradient values from a normalizing flow, which allows the perception system to optimize local regions rather than the whole image. By setting up a table top picking experiment with exceptionally difficult lighting conditions, we show that our method achieves a 60% higher success rate for an object detection task compared to previous methods.

Figures

Figures reproduced from arXiv: 2412.07565 by the authors.

**Figure 2.** Left: Images from the COCO validation set. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Illustration of how we apply NFs alongside a pretrained CNN for vision tasks. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Cluttered scene used in parameter optimization experiment. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

**Figure 5.** Figure 5: Example images captured during experiments. Top left: default parameters. Top right: Default parameters with auto-exposure and auto-whitebalance. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 16 canonical work pages

[3]

Out-of- distribution detection for adaptive computer vision,

S. Kristoffersson Lind, R. Triebel, L. Nardi, and V . Krueger, “Out-of- distribution detection for adaptive computer vision,” inImage Analysis, ser. Lecture Notes in Computer Science, R. Gade, M. Felsberg, and J.-K. Kämäräinen, Eds. Cham: Springer Nature Switzerland, 2023, pp. 311–325

work page 2023
[1]

A survey of uncertainty in deep neural networks,

J. Gawlikowski, C. Rovile Njieutcheu Tassi, M. Ali, J. Lee, M. Humt, J. Feng, A. Kruspe, R. Triebel, P. Jung, R. Roscher, M. Shahzad, W. Yang, R. Bamler, and X. X. Zhu, “A survey of uncertainty in deep neural networks,” arXiv e-prints , Jul 2021, aDS Bibcode: 2021arXiv210703342G

work page 2021
[2]

A tragic loss,

Tesla, “A tragic loss,” https://www.tesla.com/blog/tragic-loss, accessed 2023-09-15

work page 2023
[4]

Yolov4: Optimal speed and accuracy of object detection,

A. Bochkovskiy, C.-Y . Wang, and H.-Y . M. Liao, “Yolov4: Optimal speed and accuracy of object detection,” Apr 2020, arXiv:2004.10934 [cs, eess]

arXiv 2020
[5]

Faster r-cnn: Towards real-time object detection with region proposal networks,

S. Ren, K. He, R. Girshick, and J. Sun, “Faster r-cnn: Towards real-time object detection with region proposal networks,” Jan 2016, arXiv:1506.01497 [cs]

arXiv 2016
[6]

Image anomaly detection with generative adversarial networks,

L. Deecke, R. Vandermeulen, L. Ruff, S. Mandt, and M. Kloft, “Image anomaly detection with generative adversarial networks,” in Machine Learning and Knowledge Discovery in Databases , M. Berlingerio, F. Bonchi, T. Gärtner, N. Hurley, and G. Ifrim, Eds. Cham: Springer International Publishing, 2019, pp. 3–17

work page 2019
[7]

Generalized odin: Detecting out-of-distribution image without learning from out-of-distribution data,

Y .-C. Hsu, Y . Shen, H. Jin, and Z. Kira, “Generalized odin: Detecting out-of-distribution image without learning from out-of-distribution data,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , June 2020

work page 2020
[8]

Principled detection of out-of- distribution examples in neural networks,

S. Liang, Y . Li, and R. Srikant, “Principled detection of out-of- distribution examples in neural networks,” CoRR, vol. abs/1706.02690, 2017

arXiv 2017

Show all 29 references

[9]

Con- trastive training for improved out-of-distribution detection,

J. Winkens, R. Bunel, A. G. Roy, R. Stanforth, V . Natarajan, J. R. Ledsam, P. MacWilliams, P. Kohli, A. Karthikesalingam, S. Kohl, A. T. Cemgil, S. M. A. Eslami, and O. Ronneberger, “Con- trastive training for improved out-of-distribution detection,”CoRR, vol. abs/2007.05566, 2020

2007 arXiv
[10]

Likelihood ratios for out-of-distribution detection,

J. Ren, P. J. Liu, E. Fertig, J. Snoek, R. Poplin, M. Depristo, J. Dillon, and B. Lakshminarayanan, “Likelihood ratios for out-of-distribution detection,” in Advances in Neural Information Processing Systems , H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alché-Buc, E. Fox, ...

2019
[11]

Self-supervised learning for generalizable out-of-distribution detection,

S. Mohseni, M. Pitale, J. Yadawa, and Z. Wang, “Self-supervised learning for generalizable out-of-distribution detection,” Proceedings of the AAAI Conference on Artificial Intelligence , vol. 34, no. 04, pp. 5216–5223, April 2020

2020
[12]

Embedded out-of-distribution detection on an autonomous robot platform,

M. Yuhas, Y . Feng, D. J. X. Ng, Z. Rahiminasab, and A. Easwaran, “Embedded out-of-distribution detection on an autonomous robot platform,” in Proceedings of the Workshop on Design Automation for CPS and IoT , ser. Destion ’21. New York, NY , USA: Association for Computing Mac...

2021
[13]

Robustness to out-of-distribution inputs via task-aware generative uncertainty,

R. McAllister, G. Kahn, J. Clune, and S. Levine, “Robustness to out-of-distribution inputs via task-aware generative uncertainty,” in International Conference on Robotics and Automation (ICRA) , 2019, pp. 2083–2089

2019
[14]

An anomaly detection method using deep convolution neural network for vision image of robot,

Y . Du, “An anomaly detection method using deep convolution neural network for vision image of robot,” Multimedia Tools and Applica- tions, vol. 79, no. 13, pp. 9629–9642, Apr 2020

2020
[15]

Unsupervised anomaly detection for a smart autonomous robotic assistant surgeon (saras) using a deep residual autoencoder,

D. J. Samuel and F. Cuzzolin, “Unsupervised anomaly detection for a smart autonomous robotic assistant surgeon (saras) using a deep residual autoencoder,” IEEE Robotics and Automation Letters , vol. 6, no. 4, pp. 7256–7261, Oct 2021

2021
[16]

Self-calibrating anomaly and change detection for autonomous inspection robots,

S. Salimpour, J. P. Queralta, and T. Westerlund, “Self-calibrating anomaly and change detection for autonomous inspection robots,” in 2022 Sixth IEEE International Conference on Robotic Computing (IRC), Dec 2022, pp. 207––214

2022
[17]

Why Normalizing Flows Fail to Detect Out-of-Distribution Data,

P. Kirichenko, P. Izmailov, and A. G. Wilson, “Why Normalizing Flows Fail to Detect Out-of-Distribution Data,” in Advances in Neural Information Processing Systems , vol. 33. Curran Associates, Inc., 2020, pp. 20 578–20 589

2020
[18]

Normalizing Flows for Probabilistic Modeling and Inference,

G. Papamakarios, E. Nalisnick, D. J. Rezende, S. Mohamed, and B. Lakshminarayanan, “Normalizing Flows for Probabilistic Modeling and Inference,” Journal of Machine Learning Research, vol. 22, no. 57, pp. 1–64, 2021

2021
[19]

Generalized out-of-distribution detection: A survey,

J. Yang, K. Zhou, Y . Li, and Z. Liu, “Generalized out-of-distribution detection: A survey,” Aug 2022, arXiv:2110.11334 [cs]

2022 arXiv
[20]

Density estimation using real nvp,

L. Dinh, J. Sohl-Dickstein, and S. Bengio, “Density estimation using real nvp,” Feb 2017, arXiv:1605.08803 [cs, stat]. [Online]. Available: http://arxiv.org/abs/1605.08803

2017 arXiv
[21]

Glow: Generative flow with invertible 1x1 convolutions,

D. P. Kingma and P. Dhariwal, “Glow: Generative flow with invertible 1x1 convolutions,” Jul 2018, arXiv:1807.03039 [cs, stat]. [Online]. Available: http://arxiv.org/abs/1807.03039

2018 arXiv
[22]

Audet and W

C. Audet and W. Hare, Derivative-Free and Blackbox Optimization . Springer, 2017

2017
[23]

Adam: A method for stochastic optimization,

D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” 2014. [Online]. Available: https://arxiv.org/abs/1412. 6980

2014
[24]

Efficientnet: Rethinking model scaling for convolutional neural networks,

M. Tan and Q. V . Le, “Efficientnet: Rethinking model scaling for convolutional neural networks,” Sep 2020, arXiv:1905.11946 [cs, stat]

2020 arXiv
[25]

ImageNet Large Scale Visual Recognition Challenge,

O. Russakovsky, J. Deng, H. Su, J. Krause, S. Satheesh, S. Ma, Z. Huang, A. Karpathy, A. Khosla, M. Bernstein, A. C. Berg, and L. Fei-Fei, “ImageNet Large Scale Visual Recognition Challenge,” International Journal of Computer Vision (IJCV) , vol. 115, no. 3, pp. 211–252, 2015

2015
[26]

Learning multiple layers of features from tiny images,

A. Krizhevsky, G. Hinton, et al., “Learning multiple layers of features from tiny images,” 2009

2009
[27]

Reading digits in natural images with unsupervised feature learning,

Y . Netzer, T. Wang, A. Coates, A. Bissacco, B. Wu, and A. Y . Ng, “Reading digits in natural images with unsupervised feature learning,” in NIPS Workshop on Deep Learning and Unsupervised Feature Learning 2011, 2011

2011
[28]

Deep learning face attributes in the wild,

Z. Liu, P. Luo, X. Wang, and X. Tang, “Deep learning face attributes in the wild,” in Proceedings of International Conference on Computer Vision (ICCV), December 2015

2015
[29]

Microsoft coco: Common objects in context,

T.-Y . Lin, M. Maire, S. Belongie, L. Bourdev, R. Girshick, J. Hays, P. Perona, D. Ramanan, C. L. Zitnick, and P. Dollár, “Microsoft coco: Common objects in context,” Feb 2015, arXiv:1405.0312 [cs]

2015 arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [3]

Out-of- distribution detection for adaptive computer vision,

S. Kristoffersson Lind, R. Triebel, L. Nardi, and V . Krueger, “Out-of- distribution detection for adaptive computer vision,” inImage Analysis, ser. Lecture Notes in Computer Science, R. Gade, M. Felsberg, and J.-K. Kämäräinen, Eds. Cham: Springer Nature Switzerland, 2023, pp. 311–325

work page 2023

[2] [1]

A survey of uncertainty in deep neural networks,

J. Gawlikowski, C. Rovile Njieutcheu Tassi, M. Ali, J. Lee, M. Humt, J. Feng, A. Kruspe, R. Triebel, P. Jung, R. Roscher, M. Shahzad, W. Yang, R. Bamler, and X. X. Zhu, “A survey of uncertainty in deep neural networks,” arXiv e-prints , Jul 2021, aDS Bibcode: 2021arXiv210703342G

work page 2021

[3] [2]

A tragic loss,

Tesla, “A tragic loss,” https://www.tesla.com/blog/tragic-loss, accessed 2023-09-15

work page 2023

[4] [4]

Yolov4: Optimal speed and accuracy of object detection,

A. Bochkovskiy, C.-Y . Wang, and H.-Y . M. Liao, “Yolov4: Optimal speed and accuracy of object detection,” Apr 2020, arXiv:2004.10934 [cs, eess]

arXiv 2020

[5] [5]

Faster r-cnn: Towards real-time object detection with region proposal networks,

S. Ren, K. He, R. Girshick, and J. Sun, “Faster r-cnn: Towards real-time object detection with region proposal networks,” Jan 2016, arXiv:1506.01497 [cs]

arXiv 2016

[6] [6]

Image anomaly detection with generative adversarial networks,

L. Deecke, R. Vandermeulen, L. Ruff, S. Mandt, and M. Kloft, “Image anomaly detection with generative adversarial networks,” in Machine Learning and Knowledge Discovery in Databases , M. Berlingerio, F. Bonchi, T. Gärtner, N. Hurley, and G. Ifrim, Eds. Cham: Springer International Publishing, 2019, pp. 3–17

work page 2019

[7] [7]

Generalized odin: Detecting out-of-distribution image without learning from out-of-distribution data,

Y .-C. Hsu, Y . Shen, H. Jin, and Z. Kira, “Generalized odin: Detecting out-of-distribution image without learning from out-of-distribution data,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , June 2020

work page 2020

[8] [8]

Principled detection of out-of- distribution examples in neural networks,

S. Liang, Y . Li, and R. Srikant, “Principled detection of out-of- distribution examples in neural networks,” CoRR, vol. abs/1706.02690, 2017

arXiv 2017