Pith. sign in

REVIEW 4 major objections 7 minor 32 references

Improving Open-Set Semantic Segmentation in 3D Point Clouds by Conditional Channel Capacity Maximization: Preliminary Results

T0 review · 4 major / 7 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a single plug-in regularizer, 3CM, which maximizes conditional mutual information between predictions and inputs for the pseudo-unknown class, lifts reported unseen-class IoU on Stanford Indoor3D from 62.3 to 80.1…

desk verdict The 17.8-point 'unseen' IoU gain is produced by supervised post-training on the unknown classes' labels, so the paper's central open-set claim is untested. read the letter →

arxiv 2505.11521 v1 pith:GYYM4IEJ submitted 2025-05-09 cs.CV eess.SP

Wang Fang , Shirin Rahimi , Olivia Bennett , Sophie Carter , Mitra Hassani , Xu Lan , Omid Javadi , Lucas Mitchell This is my paper · ORCID
classification cs.CVeess.SP
keywords open-setsemanticsegmentation3DpointcloudconditionalmutualinformationchannelcapacityregularizationNetpseudo-unknownlabelstheory
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

The paper tries to establish that a single information-theoretic regularizer, applied to the loss of an existing point-cloud segmentation network, can make the model substantially better at labeling object classes it never saw during training. It models segmentation as a conditional Markov chain and derives a term called Conditional Channel Capacity Maximization (3CM) that encourages the encoder to keep more class-conditional information in its features. On the Stanford Indoor3D benchmark, the authors report unseen-class IoU rising from 62.3 to 80.1 with the regularizer, with no drop on known classes (85.3 vs 85.4). The same term also improves part segmentation on ShapeNet Part and classification on ModelNet40. A key caveat is that the evaluation includes a post-training phase on the labels of the supposedly unseen classes.

What carries the argument

The load-bearing object is the empirical conditional mutual information (CMI) of the segmentation model: for each class $y$, the average KL divergence between a point's softmax prediction and the class-conditional mean prediction, $\mathrm{CMI}_{\mathrm{emp}}(f) = \frac{1}{N}\sum_{y\in[C]}\sum_{x_j\in D_y} \mathrm{KL}(P_{x_j}\|Q^y_{\mathrm{emp}})$, with $Q^y_{\mathrm{emp}}$ the mean prediction over class $y$. The paper's regularizer, 3CM, maximizes this quantity for the pseudo-unknown cluster only, $\mathcal{L}_{3CM} = \frac{1}{N_{\mathrm{unknown}}}\sum_{x_j\in D_{\mathrm{unknown}}} \mathrm{KL}(P_{x_j}\|\hat{Q}^y_{\mathrm{emp}})$, updating the reference $\hat{Q}^y_{\mathrm{emp}}$ with an exponential moving average (EMA factor 0.995). This single term, added to the cross-entropy loss with weight $\lambda=0.5$, is what carries the reported open-set gains.

What would settle it

Run the same Stanford Indoor3D open-set split with 3CM but delete the 30-epoch post-training phase: if unseen-class IoU no longer rises above the 62.3 baseline, then the reported 80.1 result depends on exposure to the held-out classes' labels rather than on the regularizer itself.

Watch

Extended reading notes

Core claim

The paper's central claim is that maximizing the conditional channel capacity—the average KL divergence between per-point prediction distributions and their class-conditional mean distribution—for the unknown class is a usable, generic objective for open-set segmentation. On the author's terms, 3CM is derived from modeling the segmentation pipeline as a conditional Markov chain, and it acts as a plug-in loss term requiring neither new architecture nor generative modules. The headline evidence is a 17.8-point IoU gain on unseen classes in Stanford Indoor3D (80.1 vs 62.3) alongside unchanged known-class accuracy, with similar gains across ShapeNet Part and ModelNet40. The paper is explicit about a limitation: the method 'assumes the availability of pseudo-unknown labels during training,' and its reported protocol fine-tunes the model for 30 additional epochs using the actual labels of the classes later called unseen.

Load-bearing premise

The load-bearing premise is that labels for the unknown classes are available during training (as pseudo-labels) and, in the reported evaluation, that the model can be trained on the actual labels of those 'unseen' classes for 30 epochs; without that leakage, the claimed ability to segment truly unknown categories is not established.

Editorial extensions

If this is right

  • If 3CM works as reported, any point-cloud segmentation pipeline can obtain open-set robustness by adding one loss term, without new architecture or generative models.
  • Known-class accuracy is not traded away: the S3DIS experiment shows seen-class IoU of 85.3 with 3CM versus 85.4 without.
  • The method's sensitivity to its two hyperparameters is small in a working range: $\lambda$ between 0.1 and 0.5 keeps seen-class IoU stable (85.3–85.6), and EMA factors between 0.99 and 0.999 keep unseen-class IoU between 78.1 and 80.1; the paper reports non-convergence at $\beta=0$.
  • Because the regularizer is a loss-level term, it can be combined with other open-set techniques such as uncertainty-based pseudo-labeling or synthetic unknown generation.
  • The information-theoretic formulation gives a convergence guarantee for the alternating update when batch sampling is ignored, which the ablation's loss curves support.

Reading between the lines

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

  • The reported protocol trains for 30 extra epochs on the held-out classes' labels; a fair open-set test would measure unseen-class IoU immediately after initial training, before any post-training, and the paper does not report that number.
  • If the effect reproduces, the same CMI objective could be applied to the known classes as a regularizer for class-feature separation, offering a direct test of whether conditional channel capacity on seen classes predicts robustness on related novel classes.
  • Injecting controlled noise into the pseudo-unknown labels during training would test how much the gain depends on clean unknown-class supervision, since real pseudo-labeling is noisy.
  • The single-seed, single-split comparisons in Tables 1–3 leave open how stable the 17.8-point gap is; rerunning the same split with multiple seeds and different held-out class choices would bound the effect.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 7 minor

Summary. The paper proposes Conditional Channel Capacity Maximization (3CM), a regularizer intended to improve open-set semantic segmentation in 3D point clouds. The authors model segmentation as a conditional Markov chain and claim that maximizing a conditional mutual information term between features and predictions, conditioned on class, preserves richer class-conditional information, thereby improving detection and segmentation of previously unseen classes. The method is evaluated on ShapeNet Part, ModelNet40, and Stanford Indoor3D, with reported gains such as unseen-class IoU rising from 62.3 to 80.1 on Stanford Indoor3D while seen-class accuracy remains flat. The paper also includes ablations over the regularization weight λ and the EMA factor β. The stated central claim is that 3CM enhances the network's ability to distinguish and segment previously unseen categories without modifying the base architecture.

Significance. If the central claim were supported, an information-theoretic plug-in regularizer that improves open-set segmentation in point clouds without architectural change would be a useful contribution to the computer vision community. The paper also provides machine-checked derivations only in a weak sense: the main equations are asserted rather than proven, and the experiments are not reproducible from the manuscript (no code, no dataset splits, no detailed hyperparameters). The proposed regularizer is presented as novel, but it closely follows the conditional mutual information constrained learning framework of Yang et al. (refs. [25–27]) without a reformulation that addresses the open-set setting. The reported results do not measure open-set performance: the protocol reveals the supposed unknown classes during a second training phase, so the claimed 'unseen-class' improvements are in-distribution adaptation results, not evidence of novel-category detection. This is a load-bearing flaw that invalidates the paper's headline claims, and it cannot be fixed by minor changes within the manuscript's current scope.

major comments (4)
  1. [Section 5, Table 3] The evaluation protocol does not test open-set semantic segmentation as defined in the abstract and introduction. After initial training, the authors state they 'remove the output column corresponding to the unknown class, append 8 new output columns, and train the model for an additional 30 epochs' on the actual labels of the previously held-out classes. Equation (8) trains with CE(P_x_unknown, Y_unknown) plus the 3CM term, which requires ground-truth labels for unknown samples. The headline result of 62.3 to 80.1 unseen IoU in Table 3 is therefore a measure of how well the model can fit an expanded output head when given the held-out labels, not of detecting previously unseen categories. For a true open-set protocol, no label or output column for the target unseen classes may appear during training; here the entire second-phase training violates that requirement.
  2. [Section 4.1, Eqs. (4)–(5)] The derivation of the empirical CMI in Eq. (5) from the definition in Eqs. (1)–(4) is asserted rather than shown. The joint entropy and conditional distributions are not specified clearly: Eq. (2) contains P(Ŷ=1|x) and P_{Ŷ|y}(Ŷ=i|Y=i), which mix indices and condition on a single class y, yet Eq. (5) sums over all y∈[C] with Q^y_emp defined as the mean of P_{x_j}. The transition from a per-class conditional KL divergence to a sum over all classes with a class-specific mean distribution is nontrivial and requires an explicit derivation, especially because the conditioning variable Y is discrete and the sum over y in Eq. (5) appears to marginalize Y rather than condition on it. As written, Eq. (5) is not justified by Eq. (4).
  3. [Section 4.3, Eq. (8) and Section 7] The paper admits in Section 7 that '3CM assumes the availability of pseudo-unknown labels during training.' This assumption is not merely a limitation but is central to the method: the loss in Eq. (8) uses CE(P_x_unknown, Y_unknown) and the 3CM term is computed on D_unknown. In a genuine open-set deployment, pseudo-unknown labels are not available for novel classes at test time. The paper does not provide any mechanism to obtain such labels during training (e.g., synthetic unknown generation, out-of-distribution detection, or self-training), so the method reduces to a supervised adaptation scheme for classes whose labels are known in advance. This invalidates the claim that 3CM 'enhanc[es] the network's ability to distinguish and segment previously unseen categories' in the standard open-set sense.
  4. [Section 4 and Tables 1–2] The novelty of the regularizer is overstated. The proposed L3CM in Eq. (7) is essentially the same conditional mutual information regularizer introduced in refs. [25]–[27], which the paper cites but does not reformulate for the open-set setting. The authors do not explain how maximizing KL(P_x_unknown || Q_hat_emp) on pseudo-unknown samples differs from the previously proposed conditional mutual information maximization, nor why it would transfer to classes that are not represented in D_unknown. The experimental gains in Tables 1 and 2 are also based on the same flawed protocol: the 'unseen' classes are revealed during training, so these results do not support the claimed generalization to novel categories.
minor comments (7)
  1. [Abstract and Section 1] The abstract says 'Experimental results demonstrate effectiveness of proposed method on detecting unseen objects' but the only quantitative evidence involves classes whose labels are used in training; the wording should be revised to avoid claiming open-set detection.
  2. [Section 2] Related work discusses several prior methods, but no open-set baseline is included in the experiments. Reporting comparisons to the cited PDF [23] and APF [12] methods would be necessary to support the claim of state-of-the-art performance.
  3. [Section 5, Table 1 caption and table body] The ShapeNet Part table header is malformed: the column names and shape counts are misaligned, and some cells contain only dashes. The table is hard to read and should be reformatted.
  4. [Section 5] There are multiple grammatical errors: 'In this section, we conclude the paper by conduct a through set of experiments' and 'improvement in opem-set segmentation performance.' The paper should be proofread.
  5. [Section 5 and Figure 1] Figure 1 is described as showing Stanford-Indoor3D results, but the figure caption references Table 1 for alignment; the caption should be corrected to reference Table 3.
  6. [Section 6.1, Figure 2] Figure 2 shows the objective value over the unseen class, but the y-axis label is missing and the caption does not define what 'objective value' means. Please clarify.
  7. [Section 7] The discussion correctly lists limitations, but these limitations are not addressed in the experiments. In particular, the CMI estimation sensitivity and the assumption of pseudo-unknown labels should be tested empirically rather than deferred to future work.

Circularity Check

2 steps flagged · score 8.0 of 10

Headline unseen-IoU gain is measured after fine-tuning on the labels of the supposedly unseen classes, so the central open-set claim reduces to fitting those labels.

  1. self definitional [Section 4.2 (unknown-class definition) vs. Section 4.3 Eq. (8); Section 8 limitation]
    "a special class—often referred to as the unknown class—represents categories that are not seen during training [24] ... The modified loss function is defined as follows: min θ CE(Pxunknown,Yunknown)+ λL3CM (Pxunknown, Q̂y_emp), (8) ... Although effective, 3CM assumes the availability of pseudo-unknown labels during training, which may not always hold in practice."

    The unknown class is defined as made of categories not seen during training, but Eq. (8) trains with CE(P_x_unknown, Y_unknown) on the unknown samples, and Section 4.3 says the 3CM term is activated whenever an unknown-class sample is present in the current batch. The labels Y_unknown are thus inputs to the loss, so the quantity called 'unknown' is label-observed during training by construction. The paper's own Section 8 confirms that unknown-class labels are assumed available during training.

  2. fitted input called prediction [Section 5, ShapeNet protocol and Table 3 / Stanford-Indoor3D results]
    "We group the classes lamp, laptop, motor, mug, pistol, rocket, skateboard, and table into a single ”unknown” class. After the initial training phase, we remove the output column corresponding to the unknown class, append 8 new output columns, and train the model for an additional 30 epochs."

    Table 3's headline result (unseen IoU rising from 62.3 to 80.1 with 3CM) is evaluated after the model has been trained for 30 extra epochs on the actual labels of the eight classes it was supposed to detect as 'previously unseen.' The 17.8-point gain is therefore a fit to the held-out labels, not a measurement of open-set detection. The reported 'prediction' has the answer as a training input.

full rationale

The mathematical apparatus in Section 4 (Markov-chain model, empirical CMI) is a legitimate reformulation, and the paper's use of the CMI regularizer would not by itself be circular if the evaluation respected open-set assumptions. The circularity is in the experimental protocol: the loss in Eq. (8) needs unknown-class labels (D_unknown, Y_unknown), and Section 5 further trains for 30 epochs on the actual labels of the classes later reported as 'unseen.' Thus the central quantitative claim—that 3CM improves detection of previously unseen categories, e.g., Table 3's 62.3 -> 80.1—reduces by construction to fitting the very labels that are supposed to be withheld. The paper's own Section 8 admission that '3CM assumes the availability of pseudo-unknown labels during training' confirms the labels are part of the training input. Because the main empirical prediction is untested as an open-set claim, the score is high. The mathematical derivation itself does not reduce to its output, but the headline experimental result does.

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

The central claim rests on the availability of pseudo-unknown labels and on the identification of the empirical KL average with true conditional mutual information. The two fitted hyperparameters lambda and beta are tuned on the same benchmarks used for evaluation. No new physical or conceptual entities are introduced; the regularizer is constructed from standard information-theoretic quantities.

free parameters (4)
  • lambda = 0.5 (selected by ablation on Stanford Indoor3D, Table 4)
    Weight for the 3CM term in Eq. 8; chosen on the same dataset used for final evaluation, creating selection bias.
  • beta = 0.995 (selected by ablation, Table 5)
    EMA factor in Eq. 9; selected on the same benchmark used for the headline result.
  • second-phase training epochs = 30
    Number of epochs for fine-tuning on the unknown classes after removing the unknown output column (Section 5), chosen by hand.
  • unknown class grouping = 8 ShapeNet categories grouped as unknown
    Lamp, laptop, motor, mug, pistol, rocket, skateboard, table are grouped into one 'unknown' class (Section 5); the choice of which classes are 'unknown' affects results.
assumptions (4)
  • standard math Entropy, KL divergence, and conditional mutual information definitions from Cover apply to the network's discrete softmax outputs.
    Section 3 and 4 rely on these definitions for the regularizer.
  • domain assumption The segmentation pipeline can be modeled as a conditional Markov chain X to X-hat to Y-hat.
    Section 4.1 states this modeling without proof; the validity of the chain and the identity I(Y-hat; X | Y) = E KL(...) are assumed.
  • ad hoc to paper The empirical average in Eq. 5 is a valid estimator of conditional mutual information, and maximizing it on pseudo-unknown samples improves generalization to genuinely novel classes.
    Eq. 5 is asserted rather than derived; the transfer from pseudo-unknown to true open-set classes is assumed in Section 4.3 and not tested.
  • ad hoc to paper Pseudo-unknown labels are available during training.
    Eq. 8 uses D_unknown; the Discussion (Section 7) admits this assumption, which does not hold in true open-set settings.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Improving Open-Set Semantic Segmentation in 3D Point Clouds by Conditional Channel Capacity Maximization: Preliminary Results." pith.science (2026). https://pith.science/paper/GYYM4IEJ

@misc{pith2026250511521,
  author       = {Pith},
  title        = {Pith review of: Improving Open-Set Semantic Segmentation in 3D Point Clouds by Conditional Channel Capacity Maximization: Preliminary Results},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GYYM4IEJ}},
  note         = {Machine review of arXiv:2505.11521}
}
read the original abstract

Point-cloud semantic segmentation underpins a wide range of critical applications. Although recent deep architectures and large-scale datasets have driven impressive closed-set performance, these models struggle to recognize or properly segment objects outside their training classes. This gap has sparked interest in Open-Set Semantic Segmentation (O3S), where models must both correctly label known categories and detect novel, unseen classes. In this paper, we propose a plug and play framework for O3S. By modeling the segmentation pipeline as a conditional Markov chain, we derive a novel regularizer term dubbed Conditional Channel Capacity Maximization (3CM), that maximizes the mutual information between features and predictions conditioned on each class. When incorporated into standard loss functions, 3CM encourages the encoder to retain richer, label-dependent features, thereby enhancing the network's ability to distinguish and segment previously unseen categories. Experimental results demonstrate effectiveness of proposed method on detecting unseen objects. We further outline future directions for dynamic open-world adaptation and efficient information-theoretic estimation.

Figures

Figures reproduced from arXiv: 2505.11521 by the authors.

Figure 1
Figure 1. Semantic segmentation results on Stanford-indoor3d [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. The objective value over the unseen class, the value [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Segmentation results using varying regularization pa [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 26 canonical work pages

  1. [25]

    Conditional mutual informa- tion constrained deep learning for classification (2023).URL https://arxiv

    En-Hui Yang, Shayan Mohajer Hamidi, Linfeng Ye, Ren- hao Tan, and Beverly Yang. Conditional mutual informa- tion constrained deep learning for classification (2023).URL https://arxiv. org/abs/2309.09123. 5

    arXiv 2023

  2. [27]

    Conditional mutual information constrained deep learning for classification

    En-Hui Yang, Shayan Mohajer Hamidi, Linfeng Ye, Ren- hao Tan, and Beverly Yang. Conditional mutual information constrained deep learning for classification. IEEE Transac- tions on Neural Networks and Learning Systems , pages 1– 13, 2025. 2, 3, 4

    work page 2025

  3. [1]

    Monocular 3d object detection for autonomous driving

    Xiaozhi Chen, Kaustav Kundu, Ziyu Zhang, Huimin Ma, Sanja Fidler, and Raquel Urtasun. Monocular 3d object detection for autonomous driving. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recog- nition (CVPR), 2016. 1

    work page 2016

  4. [2]

    4d spatio-temporal convnets: Minkowski convolutional neural networks

    Christopher Choy, JunYoung Gwak, and Silvio Savarese. 4d spatio-temporal convnets: Minkowski convolutional neural networks. In Proceedings of the IEEE Conference on Com- puter Vision and Pattern Recognition , pages 3075–3084,

  5. [3]

    Elements of information theory

    Thomas M Cover. Elements of information theory . John Wiley & Sons, 1999. 3

    1999

  6. [4]

    Deep multi-modal object de- tection and semantic segmentation for autonomous driving: Datasets, methods, and challenges

    Di Feng, Christian Haase-Sch ¨utz, Lars Rosenbaum, Heinz Hertlein, Claudius Gl ¨aser, Fabian Timm, Werner Wies- beck, and Klaus Dietmayer. Deep multi-modal object de- tection and semantic segmentation for autonomous driving: Datasets, methods, and challenges. IEEE Transactions on Intelligent Transportation Systems, 22(3):1341–1360, 2021. 1

    work page 2021

  7. [5]

    Deep learning for 3d point clouds: A survey

    Yulan Guo, Hanyun Wang, Qingyong Hu, Hao Liu, Li Liu, and Mohammed Bennamoun. Deep learning for 3d point clouds: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(12):4338–4364, 2021. 1

    work page 2021

  8. [6]

    Robustness against adversarial attacks via learning confined adversarial poly- topes

    Shayan Mohajer Hamidi and Linfeng Ye. Robustness against adversarial attacks via learning confined adversarial poly- topes. In ICASSP 2024 - 2024 IEEE International Confer- ence on Acoustics, Speech and Signal Processing (ICASSP), pages 5670–5674, 2024. 3

    work page 2024

Show all 32 references
  1. [7]

    Adversarial train- ing via adaptive knowledge amalgamation of an ensemble of teachers, 2024

    Shayan Mohajer Hamidi and Linfeng Ye. Adversarial train- ing via adaptive knowledge amalgamation of an ensemble of teachers, 2024. 3

    2024

  2. [8]

    Distributed quasi- newton method for fair and fast federated learning

    Shayan Mohajer Hamidi and Linfeng Ye. Distributed quasi- newton method for fair and fast federated learning. Transac- tions on Machine Learning Research, 2025. 3

    2025

  3. [9]

    Fed-it: Addressing class imbalance in feder- ated learning through an information- theoretic lens

    Shayan Mohajer Hamidi, Renhao Tan, Linfeng Ye, and En- Hui Yang. Fed-it: Addressing class imbalance in feder- ated learning through an information- theoretic lens. In 2024 IEEE International Symposium on Information Theory (ISIT), pages 1848–1853, 2024. 5

    2024

  4. [10]

    Pointcam: Cut-and- mix for open-set point cloud learning

    Jie Hong, Shi Qiu, Weihong Li, Saeed Anwar, Mehrtash Ha- randi, Nick Barnes, and Lars Petersson. Pointcam: Cut-and- mix for open-set point cloud learning. 2022. 2

    2022

  5. [11]

    3d convolu- tional neural networks for human action recognition

    Shuiwang Ji, Wei Xu, Ming Yang, and Kai Yu. 3d convolu- tional neural networks for human action recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 35(1):221–231, 2013. 4

    2013

  6. [12]

    Open-set semantic segmenta- tion for point clouds via adversarial prototype framework

    Jianan Li and Qiulei Dong. Open-set semantic segmenta- tion for point clouds via adversarial prototype framework. In 2023 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), pages 9425–9434, 2023. 1, 2

    2023

  7. [13]

    Pointnet: Deep learning on point sets for 3d classification and segmentation

    Charles R Qi, Hao Su, Kaichun Mo, and Leonidas J Guibas. Pointnet: Deep learning on point sets for 3d classification and segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 652–660,

  8. [14]

    Qi, Li Yi, Hao Su, and Leonidas J

    Charles R. Qi, Li Yi, Hao Su, and Leonidas J. Guibas. Point- net++: deep hierarchical feature learning on point sets in a metric space. In Proceedings of the 31st International Conference on Neural Information Processing Systems, page 5105–5114, Red Hook, NY , USA, 2017. Curra...

    2017

  9. [15]

    Novel class discovery for 3d point cloud semantic segmen- tation

    Luigi Riz, Cristiano Saltori, Elisa Ricci, and Fabio Poiesi. Novel class discovery for 3d point cloud semantic segmen- tation. In 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 9393–9402, 2023. 2

    2023

  10. [16]

    Towards 3d object maps for au- tonomous household robots

    Radu Bogdan Rusu, Nico Blodow, Zoltan Marton, Alina Soos, and Michael Beetz. Towards 3d object maps for au- tonomous household robots. In2007 IEEE/RSJ International Conference on Intelligent Robots and Systems, pages 3191– 3198, 2007. 1

    2007

  11. [17]

    C. E. Shannon. A mathematical theory of communication. The Bell System Technical Journal, 27(3):379–423, 1948. 3

    1948

  12. [18]

    Kpconv: Flexible and deformable convolution for point clouds

    Hugues Thomas, Charles R Qi, Jean-Emmanuel Deschaud, Beatriz Marcotegui, Franc ¸ois Goulette, and Leonidas J Guibas. Kpconv: Flexible and deformable convolution for point clouds. In Proceedings of the IEEE/CVF international conference on computer vision, pages 6411–6420, 2019. 1, 2

    2019

  13. [19]

    Weijie Wei, Osman ¨Ulger, Fatemeh Karimi Nejadasl, Theo Gevers, and Martin R. Oswald. Auto-vocabulary segmenta- tion for lidar points, 2024. 2

    2024

  14. [20]

    Apex: Unsupervised, object-centric scene segmenta- tion and tracking for robot manipulation

    Yizhe Wu, Oiwi Parker Jones, Martin Engelcke, and Ingmar Posner. Apex: Unsupervised, object-centric scene segmenta- tion and tracking for robot manipulation. In 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 3375–3382, 2021. 1

    2021

  15. [21]

    Interactive shape co-segmentation via label propagation

    Zizhao Wu, Ruyang Shou, Yunhai Wang, and Xinguo Liu. Interactive shape co-segmentation via label propagation. Computers and Graphics, 38:248–254, 2014. 4

    2014

  16. [22]

    3d shapenets: A deep representation for volumetric shapes

    Zhirong Wu, Shuran Song, Aditya Khosla, Fisher Yu, Lin- guang Zhang, Xiaoou Tang, and Jianxiong Xiao. 3d shapenets: A deep representation for volumetric shapes. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1912–1920, 2015. 1, 5

    1912

  17. [23]

    Pdf: A probability-driven framework for open world 3d point cloud semantic segmen- tation

    Jinfeng Xu, Siyuan Yang, Xianzhi Li, Yuan Tang, Yixue Hao, Long Hu, and Min Chen. Pdf: A probability-driven framework for open world 3d point cloud semantic segmen- tation. In 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 5977–5986, 2024. 1, 2

    2024

  18. [24]

    Markov knowledge distil- lation: Make nasty teachers trained by self-undermining knowledge distillation fully distillable

    En-hui Yang and Linfeng Ye. Markov knowledge distil- lation: Make nasty teachers trained by self-undermining knowledge distillation fully distillable. In Computer Vision – ECCV 2024, pages 154–171, Cham, 2025. Springer Nature Switzerland. 3

    2024

  19. [26]

    Conditional mutual information con- strained deep learning: Framework and preliminary results

    En-Hui Yang, Shayan Mohajer Hamidi, Linfeng Ye, Renhao Tan, and Beverly Yang. Conditional mutual information con- strained deep learning: Framework and preliminary results. In 2024 IEEE International Symposium on Information The- ory (ISIT), pages 569–574, 2024. 5

    2024

  20. [28]

    Methods and systems for conditional mutual information constrained deep learning, 2025

    En-hui Yang, Shayan Mohajer Hamidi, et al. Methods and systems for conditional mutual information constrained deep learning, 2025. US Patent App. 18/829,437. 2

    2025

  21. [29]

    Bayes conditional distribution estimation for knowledge distillation based on conditional mutual informa- tion

    Linfeng Ye, Shayan Mohajer Hamidi, Renhao Tan, and EN- HUI Y ANG. Bayes conditional distribution estimation for knowledge distillation based on conditional mutual informa- tion. In The Twelfth International Conference on Learning Representations, 2024. 3

    2024

  22. [30]

    Towards undistillable models by minimizing conditional mutual information, 2025

    Linfeng Ye, Shayan Mohajer Hamidi, Renhao Tan, and EN- HUI Y ANG. Towards undistillable models by minimizing conditional mutual information, 2025. 4

    2025

  23. [31]

    Kim, Duygu Ceylan, I-Chao Shen, Mengyan Yan, Hao Su, Cewu Lu, Qixing Huang, Alla Shef- fer, and Leonidas Guibas

    Li Yi, Vladimir G. Kim, Duygu Ceylan, I-Chao Shen, Mengyan Yan, Hao Su, Cewu Lu, Qixing Huang, Alla Shef- fer, and Leonidas Guibas. A scalable active framework for region annotation in 3d shape collections. ACM Trans. Graph., 35(6), 2016. 4

    2016

  24. [32]

    Real- time 3d segmentation for human-robot interaction

    Andre ¨Uckermann, Robert Haschke, and Helge Ritter. Real- time 3d segmentation for human-robot interaction. In 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems, pages 2136–2143, 2013. 1

    2013

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.