Pith. sign in

REVIEW 5 major objections 6 minor 41 references

Few-shot Structure-Informed Machinery Part Segmentation with Foundation Models and Graph Neural Networks

T0 review · 5 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read 1–25 labeled images train a part segmenter in minutes

desk verdict A fresh pipeline with an evaluation leak: the headline few-shot numbers are not clean held-out estimates, so the paper needs a revision, not a desk reject. read the letter →

arxiv 2501.10080 v1 pith:QAZRVLWS submitted 2025-01-17 cs.CV

Michael Schwingshackl , Fabio Francisco Oberweger , Markus Murschitz This is my paper · ORCID
classification cs.CV
keywords few-shotsegmentationgraphneuralnetworksfoundationmodelsCLIPSegSegmentAnythingModelSuperPointsynthetic-to-realtransfermachinerypart
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 few-shot part segmentation for machinery does not require dense pixel labels or heavy task-specific training: it can be reduced to classifying a graph of interest points whose features are enriched by text-driven foundation models, followed by a foundation-model segmentation step. The load-bearing move is to keep CLIPSeg, SuperPoint, and SAM frozen and train only a small graph network on a handful of annotated images. The authors demonstrate this on a synthetic truck-mounted crane at five granularities, with training under five minutes and J&F up to 90.1 on synthetic data and 92.2 on real footage after synthetic-only training. The same pipeline also generalizes to DAVIS 2017 video segmentation with as few as one to three support frames. If correct, this points to a practical recipe for adapting perception systems to new machines and structures quickly, with little labeled data.

What carries the argument

The central object is an image-spanning undirected graph built from SuperPoint interest points; each node carries the SuperPoint descriptor concatenated with a sigmoid-normalized CLIPSeg likelihood for structure-related text prompts, and edges connect each node to its k nearest neighbors. A SAGEConv-based graph neural network classifies the nodes into part labels, and a prompt-engineering module (isolation forest, Mahalanobis filtering, and farthest point sampling) converts the classified nodes into point and box prompts for SAM. The graph representation is what carries the argument: it is permutation-invariant but adjacency-preserving, so articulations of the crane change pixel coordinates but not the relational structure the GCN learns from, which is what makes few-shot training and synthetic-to-real transfer feasible.

What would settle it

Take a real or rendered machine image containing a large flat, textureless part, such as a smooth hydraulic cylinder, and give the pipeline one or more annotated examples of that part. If SuperPoint detects no interest points inside the part's mask, the trainable classifier never sees nodes for that part and SAM is never prompted for it, so the predicted mask will be empty regardless of training set size.

Watch

Extended reading notes

Core claim

The paper's central claim is that composite objects such as a truck-mounted loading crane can be segmented into functional parts by classifying a sparse graph of image interest points rather than by dense pixel classification. SuperPoint supplies the graph nodes, CLIPSeg attaches text-prompt likelihoods to each node, and a small graph convolutional network trained on 1–25 annotated samples labels the nodes; the labeled nodes are then converted into point-and-box prompts that drive SAM to produce precise masks. On a synthetic crane dataset the method reaches J&F scores of 90.1, 80.4, and 55.8 for the Truck, Truck Crane, and eight-class Low granularities, and after training on ten synthetic supports it scores up to 92.2 on real crane images. The same pipeline, trained on the first frame of DAVIS 2017 videos, reaches J&F 54.5 with one frame and scores in the 71.5–74.5 range with three support frames, without using temporal information. Training takes under five minutes on consumer GPUs.

Load-bearing premise

The method assumes that SuperPoint interest points, augmented with CLIPSeg features, place at least one graph node on every part that needs to be segmented, including small and flat regions; if a part produces no interest points, the GCN never labels it and SAM is never prompted to segment it.

Editorial extensions

If this is right

  • A new part-segmentation task with up to 25 annotated images can be trained in under five minutes on consumer GPUs, and inference on a single image takes under one second.
  • Training only on synthetic renderings transfers to real camera images of the same machine type, with J&F up to 92.2 on the Truck granularity.
  • The same frozen foundation models and trainable GCN act as a general few-shot segmenter: with only the first frame of a DAVIS 2017 sequence as support it reaches J&F 54.5, and with three support frames it reaches roughly 71.5–74.5.
  • Adding CLIPSeg likelihoods to the node features improves J&F by about five points, and point-only SAM prompts outperform combined point-and-box prompts on DAVIS 2017.
  • The method outperforms a COCO-pretrained Mask R-CNN fine-tuned on 15 samples, both with and without SAM refinement, at every tested granularity.

Reading between the lines

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

  • Because only the GCN is trained, the system's ceiling is set by the frozen models' coverage; a stronger interest-point detector or a dense proposal stage added upstream should directly raise performance on fine and flat parts without changing the rest of the pipeline.
  • The DAVIS 2017 results suggest the same orchestration could be applied to other generalist few-shot tasks, such as segmenting furniture, vehicle parts, or infrastructure elements, wherever a text prompt and a few annotated images are available.
  • A testable extension would be swapping CLIPSeg's text encoder for a more capable open-vocabulary model and measuring whether the reported J&F gains on DAVIS and the crane dataset scale accordingly.
  • The synthetic-to-real result is demonstrated on one crane model; establishing the transfer more broadly would require evaluating on additional machine types and real-world clutter.
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

5 major / 6 minor

Summary. The paper proposes a few-shot machinery part segmentation pipeline that combines three frozen foundation models (CLIPSeg, SuperPoint, SAM) with a small trainable graph convolutional network (GCN). Interest points detected by SuperPoint are enhanced with CLIPSeg logit features and connected into a nearest-neighbor graph; the GCN classifies graph nodes into part labels, and the classified nodes are converted into point and bounding-box prompts for SAM to produce final pixel masks. The method is evaluated on a synthetic truck-mounted loading crane dataset at five granularity levels, on three real crane images, and on DAVIS 2017 for semi-supervised video segmentation. The paper claims effective few-shot segmentation with 1–25 synthetic samples, training under five minutes, and a J&F of 92.2 on real data with 10 synthetic support samples.

Significance. If the reported numbers were clean, the paper would make a useful practical contribution: orchestrating off-the-shelf foundation models with a light GCN for few-shot part segmentation is an appealing recipe with real relevance to autonomous systems, and the synthetic data generation pipeline is a reusable asset. The explicit reporting of hyperparameters, training times, and the availability of code and sample data (claimed in the footnote) are also positive. However, the central quantitative claims are not currently supported because hyperparameters were selected using the evaluation test set, the real-data result is a cherry-picked maximum over only three images, the DAVIS number in the abstract is inconsistent with the paper's own table, and the training-time claim is contradicted by the reported table. These issues undermine the reliability of the headline results as held-out estimates.

major comments (5)
  1. [§4.5 and §5.1] The hyperparameter search in §4.5 uses 100 test samples from the same 250-sample test set on which the final evaluation in §5.1 is performed, and the supplementary tables (e.g., Tables 8–17) state that each tuning run was evaluated on 250 test samples. Selecting hyperparameters from the test distribution before computing the reported J&F scores means the numbers in Table 3 are optimistic, not unbiased held-out estimates. This affects the paper's central few-shot claims. Please re-run the evaluation with model selection on a separate validation split that is disjoint from the test set, and report corrected numbers.
  2. [§5.1 "Synthetic to real"] The abstract and §6 cite a real-data J&F of 92.2, but this is the maximum of only three real images (for Truck: 84.0, 88.9, 92.2; for Truck Crane: 65.6, 65.1, 73.7). Reporting the best of three images as evidence of "robust generalization" is not a statistically sound summary. Please report mean ± std over the curated real set, state the selection protocol, and ideally enlarge the real test set before making generalization claims.
  3. [Abstract, §5.2, Table 4] The abstract claims "a J&F score of 71.5 in semi-supervised video segmentation with three support samples," but Table 4 reports Ours(FLM) with J&F = 74.5, J = 71.5, F = 77.6; the abstract appears to report the J component as J&F. Additionally, Table 4 mixes general-purpose methods (Painter, SegGPT) with video-specific methods (OSMN, OSVOS, STM, XMem) that use temporal information and task-specific training; without a clear separation of these settings, the comparative table can mislead readers about the significance of the result. Please correct the abstract and restrict the comparison to properly matched baselines or clearly discuss the incomparability.
  4. [§5.4, Table 6 vs. Abstract/§1] The statement that "training times are kept under five minutes" is contradicted by Table 6, which reports 434.70 s (~7.25 min) for 8 classes and 346.92 s (~5.8 min) for 22 classes. The paper's headline speed claim is therefore inaccurate for two of the five evaluated granularities. Please correct the claim or adjust the training setup (e.g., fewer epochs) so that the stated bound is met.
  5. [§4.1 and §6] The method's ability to segment every part of an object depends on SuperPoint detecting at least one interest point on each part, and the authors acknowledge in §6 that "interest points do not ensure perfect coverage on flat surfaces or in low-contrast areas." This limitation is not quantified. Please provide a per-class coverage analysis (e.g., fraction of test images where a part has no detected keypoint) and discuss how the reported scores are affected, since systematic misses could bias the evaluation upward.
minor comments (6)
  1. [General] The paper uses "GNN" and "GCN" interchangeably; please settle on one term for the trainable graph classifier and use it consistently.
  2. [Table 1 caption] There is a typo in the caption: "ganularity" should be "granularity." Similar typos appear elsewhere (e.g., "ganularity" in Table 2 caption).
  3. [Table 7 (supplementary)] In the hyperparameter description, "non-maximum-supresion" is misspelled, and the Model Type values are listed as [GAT, GAN, SAGE]; "GAN" is presumably a typo for "GCN." Please correct these.
  4. [§3] The sentence "Since the model is trained exclusively in a few-shot manner, we cannot rely on standard test-train splits" is confusing, because the paper does use separate training and test sets (100 training renderings and 250 test renderings). Please clarify the intended meaning.
  5. [Table 3 and §5.1] The baseline description says Mask R-CNN was fine-tuned on 15 training and 10 validation samples, but Table 3 labels the rows only as "25M" and "25M+S." Please clarify how these baselines relate to the 25-support-sample setting and how many support samples are actually used.
  6. [Footnote 1] The footnote promises code and sample data on GitHub but no URL is given. Please include a working URL in the final version.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the few-shot pipeline is trained on support samples and evaluated on unseen data; the main caveat (hyperparameter tuning on 100 test samples) is evaluation leakage, not circularity.

full rationale

This paper is an empirical system paper rather than a derivation chain, and no load-bearing step reduces, by the paper's own equations, to its inputs. The only trained component is a GCN whose node labels are taken from ground-truth masks of the support images (Sec. 4.2), and its few-shot scores in Table 3 are computed on 250 test samples after training on disjoint support samples, so the reported J&F numbers are not definitionally equal to the training objective. Every foundation model (CLIPSeg, SAM, SuperPoint) is used as an external, frozen, off-the-shelf module, and all citations are to independent prior work rather than to a self-citation chain; the GNN architecture comparison (GCN, GAT, SAGE) is an empirical choice, not a uniqueness theorem. The Sec. 6 admission that SuperPoint does not guarantee coverage on flat or low-contrast regions is a competence limitation and is correctly flagged by the authors; it weakens the method on fine granularities but does not make the stated results circular. The one reviewer concern that is real but outside the circularity definition is the hyperparameter search in Sec. 4.5, which was run on 100 test samples before the final evaluation on 250 test samples in Sec. 5.1; if those 100 samples are a subset of the 250, the final numbers are optimistically biased by model selection, and the paper should state the overlap or use a nested split. That is data leakage and statistical invalidity, not a reduction of the prediction to its input by construction, so it does not raise the circularity score. Accordingly, no circular steps are identified.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The method relies on hand-chosen hyperparameters, text prompts, and domain-specific assumptions about point coverage and synthetic-to-real transfer. No new physical or conceptual entities are introduced.

free parameters (3)
  • Per-granularity hyperparameter set (NR, |I|, k, MT, SP, BT, PT, SPS) = e.g., Truck: NR=4, |I|=512, k=32, MT=SAGE, SP=PB, BT=1.0, PT=1.0, SPS=20; Low: NR=4, |I|=1024, k=32, MT=SAGE, SP=PB…
    Selected by 20-run random search on 10 random train and 100 test samples (Section 4.5), and directly affect the reported J&F numbers.
  • CLIPSeg text prompts = Crane, Truck, Wheel; DAVIS sequence identifiers
    The text prompts are hand-chosen and determine the CLIPSeg likelihood maps appended to node features (Sections 4.1 and 5.3).
  • Graph augmentation hyperparameters = edge dropout, node dropout, noise levels
    Set by hand and not tuned per granularity; they influence training stability and final segmentation quality.
assumptions (5)
  • domain assumption Interest points plus CLIPSeg features are sufficient to distinguish all target parts
    The whole graph node classification depends on this; the authors note coverage limitations in Section 6.
  • domain assumption The synthetic Blender dataset is representative enough of real crane appearance for domain transfer
    The synthetic-to-real experiment (Section 5.1) assumes this; only 3 real images are used to test it.
  • domain assumption CLIPSeg's text prompts produce meaningful pixel-level likelihoods for the target classes
    Section 4.1 concatenates sigmoid CLIPSeg logits to SuperPoint descriptors; if prompts are off, features are uninformative.
  • domain assumption A k-nearest-neighbor graph on image coordinates preserves the spatial and hierarchical structure needed for part segmentation
    Section 4.1 constructs edges by Euclidean distance; this assumes proximity encodes part relations.
  • domain assumption The graph representation's permutation invariance yields lower sample complexity for part segmentation
    Invoked in Section 1 to justify small training sets, but no quantitative evidence is given for this specific task.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Few-shot Structure-Informed Machinery Part Segmentation with Foundation Models and Graph Neural Networks." pith.science (2026). https://pith.science/paper/QAZRVLWS

@misc{pith2026250110080,
  author       = {Pith},
  title        = {Pith review of: Few-shot Structure-Informed Machinery Part Segmentation with Foundation Models and Graph Neural Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QAZRVLWS}},
  note         = {Machine review of arXiv:2501.10080}
}
abstract

This paper proposes a novel approach to few-shot semantic segmentation for machinery with multiple parts that exhibit spatial and hierarchical relationships. Our method integrates the foundation models CLIPSeg and Segment Anything Model (SAM) with the interest point detector SuperPoint and a graph convolutional network (GCN) to accurately segment machinery parts. By providing 1 to 25 annotated samples, our model, evaluated on a purely synthetic dataset depicting a truck-mounted loading crane, achieves effective segmentation across various levels of detail. Training times are kept under five minutes on consumer GPUs. The model demonstrates robust generalization to real data, achieving a qualitative synthetic-to-real generalization with a $J\&F$ score of 92.2 on real data using 10 synthetic support samples. When benchmarked on the DAVIS 2017 dataset, it achieves a $J\&F$ score of 71.5 in semi-supervised video segmentation with three support samples. This method's fast training times and effective generalization to real data make it a valuable tool for autonomous systems interacting with machinery and infrastructure, and illustrate the potential of combined and orchestrated foundation models for few-shot segmentation tasks.

Figures

Figures reproduced from arXiv: 2501.10080 by the authors.

Figure 1
Figure 1. Intermediate steps of our pipeline. actions of the observed structures. For instance, the orienta￾tion of a car’s front wheel can indicate its intended direction of travel even during a standstill. Any human driver takes this into account, and so should automated machines, which are to operate safely. Moreover, a modern autonomous sys￾tem learns during operation and must adapt quickly to new functional structures. T… view at source ↗
Figure 2
Figure 2. System architecture, with all frozen foundation models (yellow) and the novel modules (green). Only the GNN is trained. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Top: The three axis of domain randomization: envi [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (17 more)
Figure 4
Figure 4. Figure 4: Five different annotation granularity levels, ranging from [PITH_FULL_IMAGE:figures/full_fig_p003_4.png]
Figure 5
Figure 5. Figure 5: Network architecture of the trainable GCN-based graph [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: Qualitative segmentation results for different support sample sizes horizontally (1 - 25) vs. structure granularity vertically. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Qualitative results, visualizing the generalization capa [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: Some qualitative DAVIS 2017 results with our 1-shot model visualizing node classification and segmentation. [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: Boxplot visualization of all hyperparameter tuning runs, demonstrating the slight but noticeable differences between various [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: Truck granularity graph node classification results for different train sample sizes. Each row shows the best five classification results for the given train sample size measured with F1 score. 8 [PITH_FULL_IMAGE:figures/full_fig_p018_10.png]
Figure 11
Figure 11. Figure 11: Truck granularity semantic segmentation results for different train sample sizes. Each row shows the best five segmentation results for the given train sample size measured with dice score. 9 [PITH_FULL_IMAGE:figures/full_fig_p019_11.png]
Figure 12
Figure 12. Figure 12: Truck Crane granularity graph node classification results for different train sample sizes. Each row shows the best five classification results for the given train sample size measured with F1 score. 10 [PITH_FULL_IMAGE:figures/full_fig_p020_12.png]
Figure 13
Figure 13. Figure 13: Truck Crane granularity semantic segmentation results for different train sample sizes. Each row shows the best five segmen￾tation results for the given train sample size measured with dice score. 11 [PITH_FULL_IMAGE:figures/full_fig_p021_13.png]
Figure 14
Figure 14. Figure 14: Low granularity graph node classification results for different train sample sizes. Each row shows the best five classification results for the given train sample size measured with F1 score. 12 [PITH_FULL_IMAGE:figures/full_fig_p022_14.png]
Figure 15
Figure 15. Figure 15: Low granularity semantic segmentation results for different train sample sizes. Each row shows the best five segmentation results for the given train sample size measured with dice score. 13 [PITH_FULL_IMAGE:figures/full_fig_p023_15.png]
Figure 16
Figure 16. Figure 16: Medium granularity graph node classification results for different train sample sizes. Each row shows the best five classification results for the given train sample size measured with F1 score. 14 [PITH_FULL_IMAGE:figures/full_fig_p024_16.png]
Figure 17
Figure 17. Figure 17: Medium granularity semantic segmentation results for different train sample sizes. Each row shows the best five segmentation results for the given train sample size measured with dice score. 15 [PITH_FULL_IMAGE:figures/full_fig_p025_17.png]
Figure 18
Figure 18. Figure 18: High granularity graph node classification results for different train sample sizes. Each row shows the best five classification results for the given train sample size measured with F1 score. 16 [PITH_FULL_IMAGE:figures/full_fig_p026_18.png]
Figure 19
Figure 19. Figure 19: High granularity semantic segmentation results for different train sample sizes. Each row shows the best five segmentation results for the given train sample size measured with dice score. 17 [PITH_FULL_IMAGE:figures/full_fig_p027_19.png]
Figure 20
Figure 20. Figure 20: Our method, trained on just 10 samples, demonstrates competitive performance when compared to fine-tuning pre-trained [PITH_FULL_IMAGE:figures/full_fig_p028_20.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 26 canonical work pages

  1. [1]

    Slic superpix- 8 els compared to state-of-the-art superpixel methods

    Radhakrishna Achanta, Appu Shaji, Kevin Smith, Aurelien Lucchi, Pascal Fua, and Sabine S ¨usstrunk. Slic superpix- 8 els compared to state-of-the-art superpixel methods. IEEE transactions on pattern analysis and machine intelligence , 34(11):2274–2282, 2012. 2

    work page 2012

  2. [2]

    Deepcut: Unsupervised segmentation using graph neu- ral networks clustering

    Amit Aflalo, Shai Bagon, Tamar Kashti, and Yonina El- dar. Deepcut: Unsupervised segmentation using graph neu- ral networks clustering. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 32–41,

    work page

  3. [3]

    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. IEEE Transactions on Pattern Anal- ysis and Machine Intelligence, 39(12):2481–2495, 2017. 2

    work page 2017

  4. [4]

    Bronstein, Joan Bruna, Taco Cohen, and Petar Velickovic

    Michael M. Bronstein, Joan Bruna, Taco Cohen, and Petar Velickovic. Geometric deep learning: Grids, groups, graphs, geodesics, and gauges. CoRR, abs/2104.13478, 2021. 2, 3

    arXiv 2021

  5. [5]

    One- shot video object segmentation

    Sergi Caelles, Kevis-Kokitsi Maninis, Jordi Pont-Tuset, Laura Leal-Taix´e, Daniel Cremers, and Luc Van Gool. One- shot video object segmentation. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 221–230, 2017. 6, 7

    work page 2017

  6. [6]

    Few shot semantic segmentation: a review of methodologies and open chal- lenges

    Nico Catalano and Matteo Matteucci. Few shot semantic segmentation: a review of methodologies and open chal- lenges. CoRR, abs/2304.05832, 2023. 2

    arXiv 2023

  7. [7]

    Xmem: Long- term video object segmentation with an atkinson-shiffrin memory model

    Ho Kei Cheng and Alexander G Schwing. Xmem: Long- term video object segmentation with an atkinson-shiffrin memory model. In European Conference on Computer Vi- sion, pages 640–658. Springer, 2022. 6, 7

    work page 2022

  8. [8]

    Superpoint: Self-supervised interest point detection and description

    Daniel DeTone, Tomasz Malisiewicz, and Andrew Rabi- novich. Superpoint: Self-supervised interest point detection and description. In Proceedings of the IEEE conference on computer vision and pattern recognition workshops , pages 224–236, 2018. 1, 3, 4

    work page 2018

Show all 41 references
  1. [9]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition ...

    2021

  2. [10]

    Fast graph representation learning with pytorch geometric

    Matthias Fey and Jan Eric Lenssen. Fast graph representation learning with pytorch geometric. CoRR, abs/1903.02428,

    1903 arXiv

  3. [11]

    Inductive representation learning on large graphs

    Will Hamilton, Zhitao Ying, and Jure Leskovec. Inductive representation learning on large graphs. Advances in neural information processing systems, 30, 2017. 4

    2017

  4. [12]

    Mask r-cnn

    Kaiming He, Georgia Gkioxari, Piotr Doll ´ar, and Ross Gir- shick. Mask r-cnn. In Proceedings of the IEEE international conference on computer vision, pages 2961–2969, 2017. 6

    2017

  5. [13]

    A generative ap- pearance model for end-to-end video object segmentation

    Joakim Johnander, Martin Danelljan, Emil Brissman, Fa- had Shahbaz Khan, and Michael Felsberg. A generative ap- pearance model for end-to-end video object segmentation. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 8953–8962, 2019. 6, 7

    2019

  6. [14]

    A review of graph neural networks: concepts, archi- tectures, techniques, challenges, datasets, applications, and future directions

    Bharti Khemani, Shruti Patil, Ketan Kotecha, and Sudeep Tanwar. A review of graph neural networks: concepts, archi- tectures, techniques, challenges, datasets, applications, and future directions. Journal of Big Data, 11(1):18, Jan 2024. 3

    2024

  7. [15]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In Yoshua Bengio and Yann LeCun, editors, 3rd International Conference on Learning Represen- tations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings, 2015. 4

    2015

  8. [16]

    Kipf and Max Welling

    Thomas N. Kipf and Max Welling. Semi-supervised classi- fication with graph convolutional networks. In 5th Interna- tional Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Pro- ceedings. OpenReview.net, 2017. 4

    2017

  9. [17]

    Segment any- thing

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment any- thing. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision, pages 4015–4026, 202...

    2023

  10. [18]

    Microsoft coco: Common objects in context

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll´ar, and C Lawrence Zitnick. Microsoft coco: Common objects in context. In Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceeding...

    2014

  11. [19]

    Isolation forest

    Fei Tony Liu, Kai Ming Ting, and Zhi-Hua Zhou. Isolation forest. In 2008 Eighth IEEE International Conference on Data Mining, pages 413–422, 2008. 5

    2008

  12. [20]

    Grounding dino: Marrying dino with grounded pre-training for open-set object detection

    Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, and Lei Zhang. Grounding dino: Marrying dino with grounded pre-training for open-set object detection. CoRR, abs/2303.05499, 2023. 3

    2023 arXiv

  13. [21]

    Part-aware prototype network for few-shot semantic segmentation

    Yongfei Liu, Xiangyi Zhang, Songyang Zhang, and Xum- ing He. Part-aware prototype network for few-shot semantic segmentation. In Computer Vision–ECCV 2020: 16th Euro- pean Conference, Glasgow, UK, August 23–28, 2020, Pro- ceedings, Part IX 16, pages 142–158. Springer, 2020. 3

    2020

  14. [22]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision, pages 10012–10022, 2021. 2

    2021

  15. [23]

    Image segmenta- tion using text and image prompts

    Timo L ¨uddecke and Alexander Ecker. Image segmenta- tion using text and image prompts. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 7086–7096, 2022. 1, 3, 4

    2022

  16. [24]

    Scaling open-vocabulary object detection

    Matthias Minderer, Alexey Gritsenko, and Neil Houlsby. Scaling open-vocabulary object detection. Advances in Neu- ral Information Processing Systems, 36, 2024. 3

    2024

  17. [25]

    On first- order meta-learning algorithms

    Alex Nichol, Joshua Achiam, and John Schulman. On first- order meta-learning algorithms. CoRR, abs/1803.02999,

    arXiv

  18. [26]

    Video object segmentation using space-time memory networks

    Seoung Wug Oh, Joon-Young Lee, Ning Xu, and Seon Joo Kim. Video object segmentation using space-time memory networks. In Proceedings of the IEEE/CVF international conference on computer vision, pages 9226–9235, 2019. 6, 7

    2019

  19. [27]

    Pytorch: An im- perative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming 9 Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An im- perative style, high-performance deep learning library. Ad- vances in neural information processing system...

    2019

  20. [28]

    The 2017 davis challenge on video object segmentation

    Jordi Pont-Tuset, Federico Perazzi, Sergi Caelles, Pablo Ar- belaez, Alexander Sorkine-Hornung, and Luc Van Gool. The 2017 davis challenge on video object segmentation. CoRR, abs/1704.00675, 2017. 6

    2017 arXiv

  21. [29]

    Explor- ing transfer learning in medical image segmentation using vision-language models

    Kanchan Poudel, Manish Dhakal, Prasiddha Bhandari, Ra- bin Adhikari, Safal Thapaliya, and Bishesh Khanal. Explor- ing transfer learning in medical image segmentation using vision-language models. CoRR, abs/2308.07706, 2023. 2

    2023 arXiv

  22. [30]

    Segment anything meets point tracking

    Frano Rajic, Lei Ke, Yu-Wing Tai, Chi-Keung Tang, Mar- tin Danelljan, and Fisher Yu. Segment anything meets point tracking. CoRR, abs/2307.01197, 2023. 3

    2023 arXiv

  23. [31]

    Grounded sam: Assembling open-world models for diverse visual tasks

    Tianhe Ren, Shilong Liu, Ailing Zeng, Jing Lin, Kun- chang Li, He Cao, Jiayu Chen, Xinyu Huang, Yukang Chen, Feng Yan, Zhaoyang Zeng, Hao Zhang, Feng Li, Jie Yang, Hongyang Li, Qing Jiang, and Lei Zhang. Grounded sam: Assembling open-world models for diverse visual tasks. CoRR...

    2024 arXiv

  24. [32]

    U- net: Convolutional networks for biomedical image segmen- tation

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

    2015

  25. [33]

    SuperGlue: Learning feature matching with graph neural networks

    Paul-Edouard Sarlin, Daniel DeTone, Tomasz Malisiewicz, and Andrew Rabinovich. SuperGlue: Learning feature matching with graph neural networks. In CVPR, 2020. 5

    2020

  26. [34]

    Metaseg: A survey of meta- learning for image segmentation

    Jiaxing Sun and Yujie Li. Metaseg: A survey of meta- learning for image segmentation. Cognitive Robotics, 1:83– 91, 2021. 2

    2021

  27. [35]

    Graph at- tention networks

    Petar Veli ˇckovi´c, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Li `o, and Yoshua Bengio. Graph at- tention networks. In International Conference on Learning Representations, 2018. 4

    2018

  28. [36]

    Ad- vances and challenges in meta-learning: A technical review

    Anna Vettoruzzo, Mohamed-Rafik Bouguelia, Joaquin Van- schoren, Thorsteinn Rognvaldsson, and KC Santosh. Ad- vances and challenges in meta-learning: A technical review. IEEE Transactions on Pattern Analysis and Machine Intelli- gence, 2024. 2

    2024

  29. [37]

    Images speak in images: A generalist painter for in-context visual learning

    Xinlong Wang, Wen Wang, Yue Cao, Chunhua Shen, and Tiejun Huang. Images speak in images: A generalist painter for in-context visual learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6830–6839, 2023. 6, 7

    2023

  30. [38]

    Seggpt: Towards seg- menting everything in context

    Xinlong Wang, Xiaosong Zhang, Yue Cao, Wen Wang, Chunhua Shen, and Tiejun Huang. Seggpt: Towards seg- menting everything in context. In ICCV, pages 1130–1140,

  31. [39]

    Hug- gingface’s transformers: State-of-the-art natural language processing

    Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chau- mond, Clement Delangue, Anthony Moi, Pierric Cistac, Tim Rault, R´emi Louf, Morgan Funtowicz, and Jamie Brew. Hug- gingface’s transformers: State-of-the-art natural language processing. CoRR, abs/1910.03771, 2019. 5

    1910 arXiv

  32. [40]

    Zonghan Wu, Shirui Pan, Fengwen Chen, Guodong Long, Chengqi Zhang, and Philip S. Yu. A comprehensive survey on graph neural networks.IEEE Transactions on Neural Net- works and Learning Systems, 32(1):4–24, 2021. 3

    2021

  33. [41]

    Efficient video object seg- mentation via network modulation

    Linjie Yang, Yanran Wang, Xuehan Xiong, Jianchao Yang, and Aggelos K Katsaggelos. Efficient video object seg- mentation via network modulation. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 6499–6507, 2018. 6, 7 10 Few-shot Structure...

    2018

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.