Pith. sign in

REVIEW 4 major objections 5 minor 40 references

ViewPCL: a point cloud based active learning method for multi-view segmentation

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read ViewPCL's Wasserstein-based inconsistency score selects superpixels whose cross-view point cloud disagreement is largest, and this policy beats ViewAL on SceneNet-RGBD.

desk verdict A clean, geometrically motivated Wasserstein inconsistency score for multi-view active learning, with an empirical claim that currently outruns the evidence: one dataset, one baseline, no error bars. read the letter →

arxiv 2506.13043 v1 pith:HNTKIW7J submitted 2025-06-16 cs.CV

Christian Hilaire , Sima Didari This is my paper · ORCID
classification cs.CV
keywords activelearningmulti-viewsegmentationsemanticuncertaintysamplingWassersteindistancesuperpixelselectionviewconsistencypointclouddistributions
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

ViewPCL proposes an active learning method for multi-view semantic segmentation that decides which superpixels an annotator should label. Its new viewpcl inconsistency score measures, with a Wasserstein distance, how much two views of the same surface disagree on where a predicted object class sits in 3D. The paper argues that this geometric disagreement signal is more informative than the pixel-wise viewpoint entropy used by ViewAL, and reports that ViewPCL achieves higher mean IoU than ViewAL on SceneNet-RGBD at the same labeling budget. A sympathetic reader should care because the method promises better data efficiency for segmentation models when annotations are scarce.

What carries the argument

The load-bearing object is the viewpcl inconsistency score s(R). For a superpixel R, the method computes overlapping subregions D where another image's class probability map cross-projects onto R, forms point cloud distributions $\mu^{q,\hat{c}_1}_1$ and $\mu^{q,\hat{c}_1}_2$ for the most probable class $\hat{c}_1$ of the reference view, and sets $s(D)$ as the average of two Wasserstein distances between these distributions (one for $\hat{c}_1$, one for the other view's most probable class $\hat{c}_2$). The superpixel score $s(R)$ is the size-weighted average of $s(D)$ over overlap subregions. The selection policy ranks superpixels by this score, then, like ViewAL, picks the cross-projected superpixel with highest view divergence to annotate.

What would settle it

An ablation on SceneNet-RGBD that computes the score over all classes instead of the top-1 class and shows no mIoU gain (or a loss) would undercut the claim that top-1 geometric disagreement is the informative signal. Alternatively, a comparison on a real indoor RGB-D dataset with noisy depth where ViewPCL fails to beat ViewAL would bound the claim to synthetic data.

Watch

Extended reading notes

Core claim

The paper's central claim is empirical: with the same number of labeled superpixels, the ViewPCL acquisition policy produces a segmentation model with higher mIoU on SceneNet-RGBD than the ViewAL policy. The score driving the policy is the viewpcl inconsistency score, which for each superpixel averages Wasserstein distances between point cloud distributions induced by class probability maps from the reference view and from cross-projected views. The method treats disagreement about the geometry of the most probable class as the uncertainty to resolve. The paper also claims this score is interpretable: it measures how badly two views disagree on the configuration of an object within a region.

Load-bearing premise

The score sums disagreement only for the single most probable class from each view, so if the most informative uncertainty lies in lower-ranked classes, the acquisition signal may miss the most valuable superpixels.

Editorial extensions

If this is right

  • If the claim holds, active learning for multi-view segmentation can be driven by cross-view geometric disagreement rather than pixel-wise entropy.
  • The same score could be applied to other dense prediction tasks with pose and depth, such as multi-view object detection or depth completion.
  • Because the score is computed from predicted class probability maps, it requires no extra labels and can be used in any round of active learning.
  • The reported improvement on SceneNet-RGBD suggests synthetic indoor datasets are a suitable testbed for view-consistency acquisition policies.
  • The method inherits ViewAL's superpixel selection and thus keeps annotation effort at the superpixel level.

Reading between the lines

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

  • The top-1 class restriction could be relaxed to a top-k or all-class Wasserstein sum; a natural test is whether multi-class disagreement further improves acquisition on scenes with many small objects.
  • One could replace the uniform selection distribution q with an uncertainty-weighted distribution, a direction the authors themselves flag as future work; this might sharpen the score in early rounds.
  • The method's reliance on accurate pose and depth means its gains may shrink on real-world data with noisy depth; testing on real indoor RGB-D scans would delineate the regime where the geometric score pays off.
  • Cross-view consistency is a strong prior; on datasets where the same surface point is rarely seen from multiple views, the score would carry little signal and random sampling might match it.
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.

Referee Report

4 major / 5 minor

Summary. The paper proposes ViewPCL, an active learning method for multi-view semantic segmentation. The method introduces a superpixel-level uncertainty score based on the Wasserstein distance between point cloud distributions induced by class probability maps from different views, aggregated over overlap regions. The selection policy combines this new 'viewpcl inconsistency score' with the view divergence score from ViewAL. The method is evaluated on SceneNet-RGBD against ViewAL using DeepLabv3+ with MobileNetV2, with 5 random seeds, and the authors claim that ViewPCL 'clearly outperforms' ViewAL in terms of mIoU at the same labeling budget.

Significance. If the empirical claim is established, ViewPCL would provide a new, geometrically interpretable uncertainty signal for multi-view active learning, with a clearly formulated score and released source code, which would be a useful contribution to the active learning literature. The method's derivation is internally consistent and does not appear circular: the score is defined a priori from model predictions and is not fitted to the outcome. However, the current evidence for the central claim is statistically weak: no error bars, no significance tests, a single baseline, a single dataset, and unexplained unequal experiment lengths. The contribution is therefore promising but not yet convincingly demonstrated.

major comments (4)
  1. [Section 5, Figure 1] The central claim that ViewPCL 'clearly outperforms ViewAL' is not supported by the reported statistics. Figure 1 shows raw and averaged mIoU curves over 5 seeds but provides no error bars, variance, confidence intervals, or significance tests. With only 5 runs, the observed advantage could lie within run-to-run variability. The authors should report per-seed paired differences between ViewPCL and ViewAL, and perform a paired statistical test (e.g., Wilcoxon signed-rank or permutation test) on the mIoU values at each round, or provide confidence intervals. Without this, the headline claim is not quantitatively grounded.
  2. [Section 5] The experiments have unequal lengths: 3 runs were executed for 7 active learning rounds and 2 for 4 rounds. The text does not explain this discrepancy, yet Figure 1b averages over all 5 runs for the first 4 rounds, mixing two different protocols. This makes the average curve potentially misleading. The authors should either run all experiments for the same number of rounds or analyze the two subgroups separately, and should state the reason for the unequal protocol.
  3. [Section 4.5 and Section 5] The evaluation uses only one baseline (ViewAL) and one dataset (SceneNet-RGBD). To support the general claim of data efficiency, the comparison should include at least a random-sampling baseline and a standard uncertainty-sampling baseline (e.g., entropy or margin), and ideally a second dataset. Furthermore, the ViewPCL selection policy in Section 3.4 combines the new viewpcl inconsistency score with ViewAL's view divergence score; without an ablation that removes one component, the observed improvement cannot be attributed to the proposed score rather than to the particular hybrid policy.
  4. [Section 3.3.2] The score in Eq. (s(D)) uses only the most probable class per view (ĉ1 and ĉ2) and a uniform selection distribution q. The authors acknowledge the flexibility of q but do not ablate these choices. If the model's top-1 class is wrong in both views, or if informative uncertainty resides in lower-ranked classes, the score may miss valuable regions. An ablation varying the number of classes considered or the selection distribution q would strengthen the claim that the proposed score captures meaningful cross-view uncertainty.
minor comments (5)
  1. [Section 4.2] The superpixel algorithm is referred to as 'SEEDs' but the standard name is 'SEEDS'. Please correct the capitalization.
  2. [References] References [12] and [13] are the same paper (Gal et al., ICML 2017). Please merge or remove the duplicate.
  3. [Section 3.3.2] The notation W in the score definition is not fully specified; the text later mentions that the Wasserstein distance is approximated by sliced Wasserstein, but it would be clearer to state this at the point of definition.
  4. [Section 5, Figure 1] The caption of Figure 1 says '4 active learning rounds' but the text explains that only 2 experiments ran exactly 4 rounds while the other 3 ran 7 rounds. Please clarify whether the figure shows only the first 4 rounds of the longer runs or only the 2 short runs, and make the caption consistent.
  5. [Appendix A] The fallback transport plan that moves µq,c1 to the boundary of the bounding box is heuristic. It would be helpful to report how often this fallback is triggered in practice, since the main text claims it occurs only in 'rare cases'.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the ViewPCL acquisition score is defined a priori from model predictions and evaluated post-hoc against ViewAL; the statistical weakness of the comparison is a reliability concern, not a circular reduction.

full rationale

ViewPCL defines its superpixel inconsistency score directly from the trained model's cross-view class probability maps: the score is the average of two Wasserstein distances between point cloud distributions induced by the top predicted classes of each view (Section 3.3.2). This score is not fitted to the downstream mIoU outcome, and no parameter is estimated from the validation labels to force the reported improvement. The comparison to ViewAL in Section 5 is an external empirical evaluation on SceneNet-RGBD, and while the absence of error bars, the unequal number of active-learning rounds, and the single baseline make the 'clearly outperforms' claim statistically fragile, this is a question of evidence strength rather than definitional equivalence. The method also does not rely on a load-bearing self-citation chain: its use of ViewAL [33] is as a baseline and as a source of standard components (superpixels, MC dropout, view divergence), not as an unverified premise that guarantees the result. Hyperparameters such as the overlap threshold and the Slice Wasserstein implementation are hand-chosen, but they are not fitted to the evaluation metric, so no fitted input is renamed as a prediction. Therefore the derivation is self-contained and no circular step can be exhibited.

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

The method introduces no new physical entities. It relies on the view-consistency assumption from ViewAL, the restriction to top-1 classes per view, and a uniform selection distribution. Several hyperparameters are left unspecified.

free parameters (4)
  • overlap_relative_size_threshold
    User-specified threshold on the relative size of overlap regions (Sections 3.3.1 and 4.2). No value or sensitivity analysis is given.
  • point_cloud_weight_threshold
    Section 4.4: points with weights below a threshold are discarded. The threshold value is not reported.
  • mc_dropout_runs_D
    Equation (1) averages over D dropout runs. D is never specified.
  • sliced_wasserstein_slices
    The number of slices for the Slice Wasserstein Distance from POT is not reported.
assumptions (4)
  • domain assumption A 3D surface point should receive consistent labels across different views
    The entire inconsistency score relies on this principle, inherited from ViewAL (Section 3 and Appendix B).
  • ad hoc to paper The most probable class per view suffices to capture cross-view uncertainty
    Section 3.3.2 restricts the score to ĉ1 and ĉ2; no ablation supports this.
  • ad hoc to paper Uniform selection distribution q on the overlap region is appropriate
    Section 3.3.2: 'we will simply set it to be the uniform distribution'; impact not analyzed.
  • domain assumption MC dropout provides a reliable approximation of predictive uncertainty
    Used in Eq (1) following ViewAL; standard but an approximation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ViewPCL: a point cloud based active learning method for multi-view segmentation." pith.science (2026). https://pith.science/paper/HNTKIW7J

@misc{pith2026250613043,
  author       = {Pith},
  title        = {Pith review of: ViewPCL: a point cloud based active learning method for multi-view segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HNTKIW7J}},
  note         = {Machine review of arXiv:2506.13043}
}
read the original abstract

We propose a novel active learning framework for multi-view semantic segmentation. This framework relies on a new score that measures the discrepancy between point cloud distributions generated from the extra geometrical information derived from the model's prediction across different views. Our approach results in a data efficient and explainable active learning method. The source code is available at https://github.com/chilai235/viewpclAL.

Figures

Figures reproduced from arXiv: 2506.13043 by the authors.

Figure 1
Figure 1. mIoU results for 4 active learning rounds. [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. mIoU results for 7 active learning rounds. [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Loss values for one of the ViewPCL experiment showing convergence of the model. [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 37 canonical work pages

  1. [1]

    Deep batch active learning by diverse, uncertain gradient lower bounds

    Jordan T Ash et al. “Deep batch active learning by diverse, uncertain gradient lower bounds”. In: arXiv preprint arXiv:1906.03671(2019)

    arXiv 2019

  2. [2]

    Segnet: A deep convolutional encoder-decoder architecture for image segmentation

    Vijay Badrinarayanan, Alex Kendall, and Roberto Cipolla. “Segnet: A deep convolutional encoder-decoder architecture for image segmentation”. In:IEEE transactions on pattern anal- ysis and machine intelligence39.12 (2017), pp. 2481–2495

    work page 2017

  3. [3]

    The power of ensembles for active learning in image classification

    William H Beluch et al. “The power of ensembles for active learning in image classification”. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2018, pp. 9368–9377

    work page 2018

  4. [4]

    Seeds: Superpixels extracted via energy-driven sampling

    Michael Van den Bergh et al. “Seeds: Superpixels extracted via energy-driven sampling”. In: European conference on computer vision. Springer. 2012, pp. 13–26

    work page 2012

  5. [5]

    Multi-View Active Learning for Video Recommendation

    Jia-Jia Cai et al. “Multi-View Active Learning for Video Recommendation.” In: IJCAI. Vol. 2019. 2019, pp. 2053–2059

    work page 2019

  6. [6]

    Encoder-decoder with atrous separable convolution for semantic imagesegmentation

    Liang-Chieh Chen et al. “Encoder-decoder with atrous separable convolution for semantic imagesegmentation”.In: Proceedings of the European conference on computer vision (ECCV). 2018, pp. 801–818

    work page 2018

  7. [7]

    Large-Scale Visual Active Learning with Deep Probabilistic Ensembles

    Kashyap Chitta, Jose M Alvarez, and Adam Lesnikowski. “Large-scale visual active learning with deep probabilistic ensembles”. In:arXiv preprint arXiv:1811.03575(2018)

    work page Pith review arXiv 2018

  8. [8]

    Multiple-Instance Active Learning

    Burr Settles Mark Craven and Soumya Ray. “Multiple-Instance Active Learning”. In: Cite- seer. 2008

    work page 2008

Show all 40 references
  1. [9]

    Committee-based sampling for training probabilistic clas- sifiers

    Ido Dagan and Sean P Engelson. “Committee-based sampling for training probabilistic clas- sifiers”. In:Machine learning proceedings 1995. Elsevier, 1995, pp. 150–157

    1995

  2. [10]

    Pot: Python optimal transport

    Rémi Flamary et al. “Pot: Python optimal transport”. In: Journal of Machine Learning Research 22.78 (2021), pp. 1–8

    2021

  3. [11]

    Selecting influential examples: Ac- tive learning with expected model output changes

    Alexander Freytag, Erik Rodner, and Joachim Denzler. “Selecting influential examples: Ac- tive learning with expected model output changes”. In:Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part IV 13. Springer. 20...

    2014

  4. [12]

    Deep bayesian active learning with image data

    Yarin Gal, Riashat Islam, and Zoubin Ghahramani. “Deep bayesian active learning with image data”. In:International conference on machine learning. PMLR. 2017, pp. 1183–1192

    2017

  5. [13]

    Deep bayesian active learning with image data

    Yarin Gal, Riashat Islam, and Zoubin Ghahramani. “Deep bayesian active learning with image data”. In: International Conference on Machine Learning. PMLR. 2017, pp. 1183– 1192

    2017

  6. [14]

    Delving deep into rectifiers: Surpassing human-level performance on imagenet classification

    Kaiming He et al. “Delving deep into rectifiers: Surpassing human-level performance on imagenet classification”. In:Proceedings of the IEEE international conference on computer vision. 2015, pp. 1026–1034. 7

    2015

  7. [15]

    Sample selection for statistical parsing

    Rebecca Hwa. “Sample selection for statistical parsing”. In: Computational linguistics 30.3 (2004), pp. 253–276

    2004

  8. [16]

    Multi-class active learning for image classification

    Ajay J Joshi, Fatih Porikli, and Nikolaos Papanikolopoulos. “Multi-class active learning for image classification”. In: 2009 ieee conference on computer vision and pattern recognition. IEEE. 2009, pp. 2372–2379

    2009

  9. [17]

    Active and continuous exploration with deep neural networks and expected model output changes

    Christoph Käding et al. “Active and continuous exploration with deep neural networks and expected model output changes”. In:arXiv preprint arXiv:1612.06129(2016)

    2016 arXiv

  10. [18]

    Batchbald: Efficient and diverse batch acquisition for deep bayesian active learning

    Andreas Kirsch, Joost Van Amersfoort, and Yarin Gal. “Batchbald: Efficient and diverse batch acquisition for deep bayesian active learning”. In: Advances in neural information processing systems32 (2019)

    2019

  11. [19]

    Fully convolutional networks for se- mantic segmentation

    Jonathan Long, Evan Shelhamer, and Trevor Darrell. “Fully convolutional networks for se- mantic segmentation”. In:Proceedings of the IEEE conference on computer vision and pattern recognition. 2015, pp. 3431–3440

    2015

  12. [20]

    Employing EM and Pool-Based Active Learning for Text Classification

    Andrew Kachites McCallum, Kamal Nigam, et al. “Employing EM and Pool-Based Active Learning for Text Classification.” In:ICML. Vol. 98. Citeseer. 1998, pp. 350–358

    1998

  13. [21]

    Scenenet rgb-d: 5m photorealistic images of synthetic indoor trajec- tories with ground truth

    John McCormac et al. “Scenenet rgb-d: 5m photorealistic images of synthetic indoor trajec- tories with ground truth”. In: (2016)

    2016

  14. [22]

    Scenenet rgb-d: Can 5m synthetic images beat generic imagenet pre-training on indoor segmentation?

    John McCormac et al. “Scenenet rgb-d: Can 5m synthetic images beat generic imagenet pre-training on indoor segmentation?” In: (2017)

    2017

  15. [23]

    Active learning with multiple views

    Ion Muslea, Steven Minton, and Craig A Knoblock. “Active learning with multiple views”. In: Journal of Artificial Intelligence Research27 (2006), pp. 203–233

    2006

  16. [24]

    Active learning using pre-clustering

    Hieu T Nguyen and Arnold Smeulders. “Active learning using pre-clustering”. In:Proceedings of the twenty-first international conference on Machine learning. 2004, p. 79

    2004

  17. [25]

    U-net: Convolutional networks for biomedicalimagesegmentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. “U-net: Convolutional networks for biomedicalimagesegmentation”.In: Medical image computing and computer-assisted intervention– MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceed- ings, pa...

    2015

  18. [26]

    Imagenet large scale visual recognition challenge

    Olga Russakovsky et al. “Imagenet large scale visual recognition challenge”. In:International journal of computer vision115.3 (2015), pp. 211–252

    2015

  19. [27]

    MobileNetV2:InvertedResidualsandLinearBottlenecks

    M.Sandleretal.“MobileNetV2:InvertedResidualsandLinearBottlenecks”.In: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition(2018), pp. 4510–4520

    2018

  20. [28]

    Optimal transport for applied mathematicians

    Filippo Santambrogio. Optimal transport for applied mathematicians. Vol. 87. Springer, 2015

    2015

  21. [29]

    Activelearningforconvolutionalneuralnetworks:Acore-set approach

    OzanSenerandSilvioSavarese.“Activelearningforconvolutionalneuralnetworks:Acore-set approach”. In:arXiv preprint arXiv:1708.00489(2017)

    2017 arXiv

  22. [30]

    Active learning literature survey

    Burr Settles. “Active learning literature survey”. In: (2009)

    2009

  23. [31]

    An Analysis of Active Learning Strategies for Sequence Labeling Tasks

    Burr Settles and Mark Craven. “An Analysis of Active Learning Strategies for Sequence Labeling Tasks”. In:Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing.Ed.byMirellaLapataandHweeTouNg.Honolulu,Hawaii:Association for Computational Linguis...

    2008

  24. [32]

    Query by committee

    H Sebastian Seung, Manfred Opper, and Haim Sompolinsky. “Query by committee”. In: Proceedings of the fifth annual workshop on Computational learning theory. 1992, pp. 287– 294

    1992

  25. [33]

    Viewal: Active learning with view- point entropy for semantic segmentation

    Yawar Siddiqui, Julien Valentin, and Matthias Nießner. “Viewal: Active learning with view- point entropy for semantic segmentation”. In:Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2020, pp. 9433–9443

    2020

  26. [34]

    Active learning for se- mantic segmentation with expected change

    Alexander Vezhnevets, Joachim M Buhmann, and Vittorio Ferrari. “Active learning for se- mantic segmentation with expected change”. In:2012 IEEE conference on computer vision and pattern recognition. IEEE. 2012, pp. 3162–3169

    2012

  27. [35]

    Cost-effective active learning for deep image classification

    Keze Wang et al. “Cost-effective active learning for deep image classification”. In: IEEE Transactions on Circuits and Systems for Video Technology27.12 (2016), pp. 2591–2600. 8

    2016

  28. [36]

    Incorporating diversity and density in active learn- ing for relevance feedback

    Zuobing Xu, Ram Akella, and Yi Zhang. “Incorporating diversity and density in active learn- ing for relevance feedback”. In:Advances in Information Retrieval: 29th European Conference on IR Research, ECIR 2007, Rome, Italy, April 2-5, 2007. Proceedings 29. Springer. 2007, pp. 246–257

    2007

  29. [37]

    Suggestive annotation: A deep active learning framework for biomedical image segmentation

    Lin Yang et al. “Suggestive annotation: A deep active learning framework for biomedical image segmentation”. In: Medical Image Computing and Computer Assisted Intervention- MICCAI 2017: 20th International Conference, Quebec City, QC, Canada, September 11-13, 2017, Proceedings,...

    2017

  30. [38]

    Learning loss for active learning

    Donggeun Yoo and In So Kweon. “Learning loss for active learning”. In:Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2019, pp. 93–102

    2019

  31. [39]

    Multi-view multi-label active learning for image classification

    Xiaoyu Zhang et al. “Multi-view multi-label active learning for image classification”. In:2009 IEEE International Conference on Multimedia and Expo. IEEE. 2009, pp. 258–261

    2009

  32. [40]

    Pyramid scene parsing network

    Hengshuang Zhao et al. “Pyramid scene parsing network”. In:Proceedings of the IEEE con- ference on computer vision and pattern recognition. 2017, pp. 2881–2890. Appendix A Computing W (µq,c 1 , µq,c 2 ) when pq 2(c) = 0 Suppose we have a subregionD of a superpixel, a selection...

    2017

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.