mlsanityData to 7 Oct 2026

No Time to Train: Empowering Non-Parametric Networks for Few-shot 3D Scene Segmentation

arXiv 2404.04050 ↗First posted 5 Apr 2024cs.CVcited by 18

Not on the map. Why?

Abstract (arXiv)

To reduce the reliance on large-scale datasets, recent works in 3D segmentation resort to few-shot learning. Current 3D few-shot segmentation methods first pre-train models on 'seen' classes, and then evaluate their generalization performance on 'unseen' classes. However, the prior pre-training stage not only introduces excessive time overhead but also incurs a significant domain gap on 'unseen' classes. To tackle these issues, we propose a Non-parametric Network for few-shot 3D Segmentation, Seg-NN, and its Parametric variant, Seg-PN. Without training, Seg-NN extracts dense representations by hand-crafted filters and achieves comparable performance to existing parametric models. Due to the elimination of pre-training, Seg-NN can alleviate the domain gap issue and save a substantial amount of time. Based on Seg-NN, Seg-PN only requires training a lightweight QUEry-Support Transferring (QUEST) module, which enhances the interaction between the support set and query set. Experiments suggest that Seg-PN outperforms previous state-of-the-art method by +4.19% and +7.71% mIoU on S3DIS and ScanNet datasets respectively, while reducing training time by -90%, indicating its effectiveness and efficiency.

Key facts. Cited by 18 papers in the map (data through 7 Oct 2026). Builds on 49 of its 59 references that are in the map. First posted on arXiv 5 Apr 2024.

Builds on

49 of 59 references.

References · 49

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  3. 2023Segment Anythingcited by 8,158
  4. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
  5. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  6. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,384
  7. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,088
  8. 2016Matching Networks for One Shot Learningcited by 2,246
  9. 2018Dynamic Graph CNN for Learning on Point Cloudscited by 2,220
  10. 2020Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domainscited by 2,170
  11. 2021Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fieldscited by 1,957
  12. 2020Point Transformercited by 1,462
37 more
  1. 2018PointCNN: Convolution On X-Transformed Pointscited by 764
  2. 2016Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modelingcited by 558
  3. 2019Revisiting Point Cloud Classification: A New Benchmark Dataset and Classification Model on Real-World Datacited by 532
  4. 2022PointNeXt: Revisiting PointNet++ with Improved Training and Scaling Strategiescited by 432
  5. 2021Tip-Adapter: Training-free CLIP-Adapter for Better Vision-Language Modelingcited by 429
  6. 2022Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Frameworkcited by 376
  7. 2017One-Shot Learning for Semantic Segmentationcited by 337
  8. 2017Few-Shot Learning with Graph Neural Networkscited by 324
  9. 2022Point-M2AE: Multi-scale Masked Autoencoders for Hierarchical Point Cloud Pre-trainingcited by 263
  10. 2017Neural Style Transfer: A Reviewcited by 199
  11. 2023Prompt, Generate, then Cache: Cascade of Foundation Models makes Strong Few-shot Learnerscited by 152
  12. 2022Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoderscited by 143
  13. 2020Pillar-based Object Detection for Autonomous Drivingcited by 137
  14. 2023Transformer-Based Visual Segmentation: A Surveycited by 124
  15. 2022Learning What Not to Segment: A New Perspective on Few-Shot Segmentationcited by 120
  16. 2022PartSLIP: Low-Shot Part Segmentation for 3D Point Clouds via Pretrained Image-Language Modelscited by 117
  17. 2023Swin3D: A Pretrained Transformer Backbone for 3D Indoor Scene Understandingcited by 111
  18. 2020Few-shot 3D Point Cloud Semantic Segmentationcited by 77
  19. 2020Self-Supervised Few-Shot Learning on Point Cloudscited by 66
  20. 2018A LiDAR Point Cloud Generator: from a Virtual World to Autonomous Drivingcited by 54
  21. 2021A Review of Location Encoding for GeoAI: Methods and Applicationscited by 53
  22. 2023PiMAE: Point Cloud and Image Interactive Masked Autoencoders for 3D Object Detectioncited by 53
  23. 2023Joint-MAE: 2D-3D Joint Masked Autoencoders for 3D Point Cloud Pre-trainingcited by 52
  24. 2019Theory of the Frequency Principle for General Deep Neural Networkscited by 45
  25. 2023Parameter is Not All You Need: Starting from Non-Parametric Networks for 3D Point Cloud Analysiscited by 34
  26. 2019FuseSeg: LiDAR Point Cloud Segmentation Fusing Multi-Modal Datacited by 33
  27. 2023Prototype Adaption and Projection for Few- and Zero-shot 3D Point Cloud Semantic Segmentationcited by 26
  28. 2022Bidirectional Feature Globalization for Few-shot Semantic Segmentation of 3D Point Cloud Scenescited by 23
  29. 2019Explicitizing an Implicit Bias of the Frequency Principle in Two-layer Neural Networkscited by 20
  30. 2023Boosting Few-shot 3D Point Cloud Segmentation via Query-Guided Enhancementcited by 19
  31. 2023Few-Shot 3D Point Cloud Semantic Segmentation via Stratified Class-Specific Attention Based Transformer Networkcited by 16
  32. 2018Edge and Corner Detection for Unorganized 3D Point Clouds with Application to Robotic Weldingcited by 14
  33. 2023Generalized Few-Shot Point Cloud Segmentation Via Geometric Wordscited by 12
  34. 2023Few-Shot Point Cloud Semantic Segmentation via Contrastive Self-Supervision and Multi-Resolution Attentioncited by 8
  35. 2023Nearest Neighbors Meet Deep Neural Networks for Point Cloud Analysiscited by 5
  36. 2023Robust Point Cloud Processing through Positional Embeddingcited by 5
  37. 2022Primitive3D: 3D Object Dataset Synthesis from Randomly Assembled Primitivescited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 5; 2025: 7; 2026: 6.

18 papers.

Sort:

By area:

  1. 2024Lift3D Foundation Policy: Lifting 2D Large-Scale Pretrained Models for Robust 3D Robotic Manipulationcited by 45
  2. 2024Any2Point: Empowering Any-modality Large Models for Efficient 3D Understandingcited by 23
  3. 2025MLA: A Multisensory Language-Action Model for Multimodal Understanding and Forecasting in Robotic Manipulationcited by 22
  4. 2024Multimodality Helps Few-shot 3D Point Cloud Semantic Segmentationcited by 20
  5. 2025Generalized Few-shot 3D Point Cloud Segmentation with Vision-Language Modelcited by 17
  6. 2024Unlearnable 3D Point Clouds: Class-wise Transformation Is All You Needcited by 16
  7. 2025Taylor Series-Inspired Local Structure Fitting Network for Few-shot Point Cloud Semantic Segmentationcited by 14
  8. 2024MM-Mixing: Multi-Modal Mixing Alignment for 3D Understandingcited by 1
  9. 2025EPSegFZ: Efficient Point Cloud Semantic Segmentation for Few- and Zero-Shot Scenarios with Language Guidancecited by 1
  10. 2026SCOPE: Scene-Contextualized Incremental Few-Shot 3D Segmentationcited by 1
  11. 2026SRENet: Spectral Re-Entry Network for Point Cloud Action Recognitioncited by 1
  12. 2025Open-world Point Cloud Semantic Segmentation: A Human-in-the-loop Frameworkcited by 0
6 more
  1. 2025Few to Big: Prototype Expansion Network via Diffusion Learner for Point Cloud Few-shot Semantic Segmentationcited by 0
  2. 2025White Aggregation and Restoration for Few-shot 3D Point Cloud Semantic Segmentationcited by 0
  3. 2026TSNN: A Non-parametric and Interpretable Framework for Traffic Time Series Forecastingcited by 0
  4. 2026A Unified Non-Parametric and Interpretable Point Cloud Analysis via t-FCW Graph Representationcited by 0
  5. 2026Training-Free Open-Vocabulary 3D Point-Cloud Segmentation on the Generalized Few-Shot Benchmarkcited by 0
  6. 2026Refine Then Fusion: Training-Free 3D Point Cloud Adaptation with Priority Refinement and Multi-Modal Knowledge Fusioncited by 0