mlsanityData to 7 Oct 2026

TeFF: Tracking-enhanced Forgetting-free Few-shot 3D LiDAR Semantic Segmentation

arXiv 2408.15657 ↗First posted 28 Aug 2024cs.CV cs.ROcited by 0

Not on the map. Why?

Abstract (arXiv)

In autonomous driving, 3D LiDAR plays a crucial role in understanding the vehicle's surroundings. However, the newly emerged, unannotated objects presents few-shot learning problem for semantic segmentation. This paper addresses the limitations of current few-shot semantic segmentation by exploiting the temporal continuity of LiDAR data. Employing a tracking model to generate pseudo-ground-truths from a sequence of LiDAR frames, our method significantly augments the dataset, enhancing the model's ability to learn on novel classes. However, this approach introduces a data imbalance biased to novel data that presents a new challenge of catastrophic forgetting. To mitigate this, we incorporate LoRA, a technique that reduces the number of trainable parameters, thereby preserving the model's performance on base classes while improving its adaptability to novel classes. This work represents a significant step forward in few-shot 3D LiDAR semantic segmentation for autonomous driving. Our code is available at https://github.com/junbao-zhou/Track-no-forgetting.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 49 of its 57 references that are in the map. First posted on arXiv 28 Aug 2024.

Builds on

49 of 57 references.

References · 49

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2020Language Models are Few-Shot Learnerscited by 32,199
  3. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  4. 2017Graph Attention Networkscited by 8,854
  5. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  6. 2016Overcoming catastrophic forgetting in neural networkscited by 4,718
  7. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,384
  8. 2016Learning without Forgettingcited by 2,465
  9. 2018Dynamic Graph CNN for Learning on Point Cloudscited by 2,220
  10. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,202
  11. 2019A Comprehensive Survey on Transfer Learningcited by 1,154
  12. 2020Center-based 3D Object Detection and Trackingcited by 1,143
37 more
  1. 20173D Semantic Segmentation with Submanifold Sparse Convolutional Networkscited by 746
  2. 2018PointConv: Deep Convolutional Networks on 3D Point Cloudscited by 656
  3. 2019Video Object Segmentation using Space-Time Memory Networkscited by 436
  4. 2022XMem: Long-Term Video Object Segmentation with an Atkinson-Shiffrin Memory Modelcited by 436
  5. 2020Searching Efficient 3D Architectures with Sparse Point-Voxel Convolutioncited by 422
  6. 2019Tracking without bells and whistlescited by 382
  7. 2020PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentationcited by 274
  8. 2017SqueezeSeg: Convolutional Neural Nets with Recurrent CRF for Real-Time Road-Object Segmentation from 3D LiDAR Point Cloudcited by 231
  9. 2018SqueezeSegV2: Improved Model Structure and Unsupervised Domain Adaptation for Road-Object Segmentation from a LiDAR Point Cloudcited by 228
  10. 2020SalsaNext: Fast, Uncertainty-aware Semantic Segmentation of LiDAR Point Clouds for Autonomous Drivingcited by 226
  11. 2021Associating Objects with Transformers for Video Object Segmentationcited by 225
  12. 2021Rethinking Space-Time Networks with Improved Memory Coverage for Efficient Video Object Segmentationcited by 200
  13. 2017SEGCloud: Semantic Segmentation of 3D Point Cloudscited by 178
  14. 2019FEELVOS: Fast End-to-End Embedding Learning for Video Object Segmentationcited by 178
  15. 2020OccuSeg: Occupancy-aware 3D Instance Segmentationcited by 177
  16. 2022Decoupling Features in Hierarchical Propagation for Video Object Segmentationcited by 146
  17. 2020Collaborative Video Object Segmentation by Foreground-Background Integrationcited by 142
  18. 2017Detect to Track and Track to Detectcited by 131
  19. 2023Rethinking Range View Representation for LiDAR Segmentationcited by 126
  20. 2020Deep Learning for LiDAR Point Clouds in Autonomous Driving: A Reviewcited by 114
  21. 2020Cylinder3D: An Effective 3D Framework for Driving-scene LiDAR Semantic Segmentationcited by 112
  22. 2021SimpleTrack: Understanding and Rethinking 3D Multi-object Trackingcited by 102
  23. 2018Efficient Video Object Segmentation via Network Modulationcited by 98
  24. 2018VideoMatch: Matching based Video Object Segmentationcited by 90
  25. 2021Efficient Regional Memory Network for Video Object Segmentationcited by 86
  26. 2021EagerMOT: 3D Multi-Object Tracking via Sensor Fusioncited by 83
  27. 2020Few-shot 3D Point Cloud Semantic Segmentationcited by 77
  28. 2018Blazingly Fast Video Object Segmentation with Pixel-Wise Metric Learningcited by 74
  29. 2020Generalized Few-shot Semantic Segmentationcited by 59
  30. 2019SalsaNet: Fast Road and Vehicle Segmentation in LiDAR Point Clouds for Autonomous Drivingcited by 57
  31. 2022CAMO-MOT: Combined Appearance-Motion Optimization for 3D Multi-Object Tracking with Camera-LiDAR Fusioncited by 50
  32. 2023Poly-MOT: A Polyhedral Framework For 3D Multi-Object Trackingcited by 32
  33. 2021Joint Multi-Object Detection and Tracking with Camera-LiDAR Fusion for Autonomous Drivingcited by 26
  34. 2020Multiple Object Tracking by Flowing and Fusingcited by 17
  35. 2020Compositional Prototype Network with Multi-view Comparision for Few-Shot Point Cloud Semantic Segmentationcited by 12
  36. 2021Generalized Few-Shot Semantic Segmentation: All You Need is Fine-Tuningcited by 8
  37. 2023Few-shot 3D LiDAR Semantic Segmentation for Autonomous Drivingcited by 3

Built on by

Not cited yet.