mlsanityData to 14 Oct 2026

CasPoinTr: Point Cloud Completion with Cascaded Networks and Knowledge Distillation

arXiv 2509.23375 ↗First posted 27 Sep 2025cs.CVcited by 0

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

Abstract (arXiv)

Point clouds collected from real-world environments are often incomplete due to factors such as limited sensor resolution, single viewpoints, occlusions, and noise. These challenges make point cloud completion essential for various applications. A key difficulty in this task is predicting the overall shape and reconstructing missing regions from highly incomplete point clouds. To address this, we introduce CasPoinTr, a novel point cloud completion framework using cascaded networks and knowledge distillation. CasPoinTr decomposes the completion task into two synergistic stages: Shape Reconstruction, which generates auxiliary information, and Fused Completion, which leverages this information alongside knowledge distillation to generate the final output. Through knowledge distillation, a teacher model trained on denser point clouds transfers incomplete-complete associative knowledge to the student model, enhancing its ability to estimate the overall shape and predict missing regions. Together, the cascaded networks and knowledge distillation enhance the model's ability to capture global shape context while refining local details, effectively bridging the gap between incomplete inputs and complete targets. Experiments on ShapeNet-55 under different difficulty settings demonstrate that CasPoinTr outperforms existing methods in shape recovery and detail preservation, highlighting the effectiveness of our cascaded structure and distillation strategy.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Detection and dense prediction distillation, one of 12 directions in Knowledge distillation. Builds on 29 of its 37 references that are in the map. First posted on arXiv 27 Sep 2025.

Builds on

29 of 37 references.

Closest ancestors · 1

Same direction, newest first

  1. 2024Make a Strong Teacher with Label Assistance: A Novel Knowledge Distillation Approach for Semantic Segmentation18 Jul

Other references · 28

  1. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,248
  2. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,401
  3. 2019Improved Knowledge Distillation via Teacher Assistantcited by 573
  4. 2018PCN: Point Completion Networkcited by 384
  5. 2019Point-Voxel CNN for Efficient 3D Deep Learningcited by 343
  6. 2021PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformerscited by 286
  7. 2020GRNet: Gridding Residual Network for Dense Point Cloud Completioncited by 174
  8. 2021SnowflakeNet: Point Cloud Completion by Snowflake Point Deconvolution with Skip-Transformercited by 164
  9. 2020Channel-wise Knowledge Distillation for Dense Predictioncited by 163
  10. 20222DPASS: 2D Priors Assisted Semantic Segmentation on LiDAR Point Cloudscited by 158
  11. 2020PF-Net: Point Fractal Network for 3D Point Cloud Completioncited by 146
  12. 2020Cross-Layer Distillation with Semantic Calibrationcited by 143
16 more
  1. 2020Cascaded Refinement Network for Point Cloud Completioncited by 112
  2. 2019Knowledge Adaptation for Efficient Semantic Segmentationcited by 105
  3. 2021Variational Relational Point Completion Networkcited by 104
  4. 2022SeedFormer: Patch Seeds based Point Cloud Completion with Upsample Transformercited by 83
  5. 2018RGCNN: Regularized Graph CNN for Point Cloud Segmentationcited by 70
  6. 2023AdaPoinTr: Diverse Point Cloud Completion with Adaptive Geometry-Aware Transformerscited by 67
  7. 2019RL-GAN-Net: A Reinforcement Learning Agent Controlled GAN Network for Real-Time Point Cloud Shape Completioncited by 58
  8. 2022PMP-Net++: Point Cloud Completion by Transformer-Enhanced Multi-step Point Moving Pathscited by 57
  9. 20213D-to-2D Distillation for Indoor Scene Parsingcited by 48
  10. 2023P2C: Self-Supervised Point Cloud Completion from Single Partial Cloudscited by 20
  11. 2022Drive&Segment: Unsupervised Semantic Segmentation of Urban Scenes via Cross-modal Distillationcited by 19
  12. 2023CFR-ICL: Cascade-Forward Refinement with Iterative Click Loss for Interactive Image Segmentationcited by 17
  13. 2022Prototypical VoteNet for Few-Shot 3D Point Cloud Object Detectioncited by 14
  14. 2023PointSSC: A Cooperative Vehicle-Infrastructure Point Cloud Benchmark for Semantic Scene Completioncited by 14
  15. 2022CP3: Unifying Point Cloud Completion by Pretrain-Prompt-Predict Paradigmcited by 9
  16. 2022Reconstruction-Aware Prior Distillation for Semi-supervised Point Cloud Completioncited by 5

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Imagine with the Teacher: Complete Shape in a Multi-View Distillation Way
  2. 2024PCDreamer: Point Cloud Completion Through Multi-view Diffusion Priors
  3. 2021A Conditional Point Diffusion-Refinement Paradigm for 3D Point Cloud Completion
  4. 20243DMambaComplete: Exploring Structured State Space Model for Point Cloud Completion
  5. 2024Position-aware Guided Point Cloud Completion with CLIP Model
  6. 2023Bidirectional Knowledge Reconfiguration for Lightweight Point Cloud Analysis
  7. 2025Self-Supervised Point Cloud Completion based on Multi-View Augmentations of Single Partial Point Cloud
  8. 2024RealDiff: Real-world 3D Shape Completion using Self-Supervised Diffusion Models