mlsanityData to 14 Oct 2026

Continual Learning for LiDAR Semantic Segmentation: Class-Incremental and Coarse-to-Fine strategies on Sparse Data

arXiv 2304.03980 ↗First posted 8 Apr 2023cs.CVcited by 9

On the map: Knowledge distillation → Detection and dense prediction distillation

Abstract (arXiv)

During the last few years, continual learning (CL) strategies for image classification and segmentation have been widely investigated designing innovative solutions to tackle catastrophic forgetting, like knowledge distillation and self-inpainting. However, the application of continual learning paradigms to point clouds is still unexplored and investigation is required, especially using architectures that capture the sparsity and uneven distribution of LiDAR data. The current paper analyzes the problem of class incremental learning applied to point cloud semantic segmentation, comparing approaches and state-of-the-art architectures. To the best of our knowledge, this is the first example of class-incremental continual learning for LiDAR point cloud semantic segmentation. Different CL strategies were adapted to LiDAR point clouds and tested, tackling both classic fine-tuning scenarios and the Coarse-to-Fine learning paradigm. The framework has been evaluated through two different architectures on SemanticKITTI, obtaining results in line with state-of-the-art CL strategies and standard offline learning.

Key facts. Cited by 9 papers in the map (data through 14 Oct 2026). Placed in Detection and dense prediction distillation, one of 12 directions in Knowledge distillation. Builds on 38 of its 53 references that are in the map. First posted on arXiv 8 Apr 2023.

Builds on

38 of 53 references.

Closest ancestors · 1

Same direction, newest first

  1. 2022Point-to-Voxel Knowledge Distillation for LiDAR Semantic Segmentation5 Jun

Other references · 37

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  2. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  3. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,248
  4. 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,859
  5. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
  6. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,401
  7. 2016Learning without Forgettingcited by 2,496
  8. 2016iCaRL: Incremental Classifier and Representation Learningcited by 2,333
  9. 2018PointPillars: Fast Encoders for Object Detection from Point Cloudscited by 1,715
  10. 2020PCT: Point cloud transformercited by 732
  11. 2020Prototypical Contrastive Learning of Unsupervised Representationscited by 640
  12. 2019RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Cloudscited by 585
25 more
  1. 2020Cylindrical and Asymmetrical 3D Convolution Networks for LiDAR Segmentationcited by 389
  2. 2021Towards Open World Object Detectioncited by 297
  3. 2020PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentationcited by 275
  4. 2020SalsaNext: Fast, Uncertainty-aware Semantic Segmentation of LiDAR Point Clouds for Autonomous Drivingcited by 226
  5. 2017Incremental Learning of Object Detectors without Catastrophic Forgettingcited by 208
  6. 2016Driving in the Matrix: Can Virtual Worlds Replace Human-Generated Annotations for Real World Tasks?cited by 194
  7. 2020Modeling the Background for Incremental Learning in Semantic Segmentationcited by 190
  8. 2017SEGCloud: Semantic Segmentation of 3D Point Cloudscited by 179
  9. 2020PLOP: Learning without Forgetting for Continual Semantic Segmentationcited by 177
  10. 2019Incremental Learning Techniques for Semantic Segmentationcited by 141
  11. 2021Continual Semantic Segmentation via Repulsion-Attraction of Sparse and Disentangled Latent Representationscited by 96
  12. 2021RECALL: Replay-based Continual Learning in Semantic Segmentationcited by 83
  13. 2022PSTNet: Point Spatio-Temporal Convolution on Point Cloud Sequencescited by 68
  14. 2020Faster ILOD: Incremental Learning for Object Detectors based on Faster RCNNcited by 61
  15. 2021Cylindrical and Asymmetrical 3D Convolution Networks for LiDAR-based Perceptioncited by 57
  16. 2021Incremental Learning in Semantic Segmentation from Image Labelscited by 39
  17. 2020Class-Incremental Learning for Semantic Segmentation Re-Using Neither Old Data Nor Old Labelscited by 31
  18. 2022Open-world Semantic Segmentation for LIDAR Point Cloudscited by 29
  19. 2022InCloud: Incremental Learning for Point Cloud Place Recognitioncited by 22
  20. 2020I3DOL: Incremental 3D Object Learning without Catastrophic Forgettingcited by 20
  21. 2022Few-shot Class-incremental Learning for 3D Point Cloud Objectscited by 13
  22. 2021Coarse-To-Fine Incremental Few-Shot Learningcited by 9
  23. 2022Learning with Style: Continual Semantic Segmentation Across Tasks and Domainscited by 9
  24. 2022Continual Coarse-to-Fine Domain Adaptation in Semantic Segmentationcited by 6
  25. 2023Learning from Mistakes: Self-Regularizing Hierarchical Representations in Point Cloud Semantic Segmentationcited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 2; 2024: 4; 2025: 1; 2026: 2.

9 papers.

  1. 2023A Survey on Continual Semantic Segmentation: Theory, Challenge, Method and Applicationcited by 21
  2. 2023Continual Road-Scene Semantic Segmentation via Feature-Aligned Symmetric Multi-Modal Networkcited by 6
  3. 2024Region-Based Optimization in Continual Learning for Audio Deepfake Detectioncited by 5
  4. 2024Is Continual Learning Ready for Real-world Challenges?cited by 4
  5. 2024MIRACLE3D: Memory-efficient Integrated Robust Approach for Continual Learning on Point Clouds via Shape Model Constructioncited by 3
  6. 2025Decoupling Continual Semantic Segmentationcited by 2
  7. 2024Enhanced Model Robustness to Input Corruptions by Per-corruption Adaptation of Normalization Statisticscited by 1
  8. 2026RoAD Benchmark: How LiDAR Models Fail under Coupled Domain Shifts and Label Evolutioncited by 0
  9. 2026Code Division Modulation Layers Against Forgetting and Inference in Continual Gait Identificationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2022Continual Learning for Class- and Domain-Incremental Semantic Segmentation
  2. 2025Improving Multimodal Distillation for 3D Semantic Segmentation under Domain Shift
  3. 2022The CLEAR Benchmark: Continual LEArning on Real-World Imagery
  4. 2023CCL: Continual Contrastive Learning for LiDAR Place Recognition
  5. 2022Continual Attentive Fusion for Incremental Learning in Semantic Segmentation
  6. 2026Continual Vision-Language Learning for Remote Sensing: Benchmarking and Analysis
  7. 2025LOSC: LiDAR Open-voc Segmentation Consolidator
  8. 2021Recent Advances of Continual Learning in Computer Vision: An Overview