mlsanityData to 7 Oct 2026

Pose Adaptive Dual Mixup for Few-Shot Single-View 3D Reconstruction

arXiv 2112.12484 ↗First posted 23 Dec 2021cs.CVcited by 5

Not on the map. Why?

Abstract (arXiv)

We present a pose adaptive few-shot learning procedure and a two-stage data interpolation regularization, termed Pose Adaptive Dual Mixup (PADMix), for single-image 3D reconstruction. While augmentations via interpolating feature-label pairs are effective in classification tasks, they fall short in shape predictions potentially due to inconsistencies between interpolated products of two images and volumes when rendering viewpoints are unknown. PADMix targets this issue with two sets of mixup procedures performed sequentially. We first perform an input mixup which, combined with a pose adaptive learning procedure, is helpful in learning 2D feature extraction and pose adaptive latent encoding. The stagewise training allows us to build upon the pose invariant representations to perform a follow-up latent mixup under one-to-one correspondences between features and ground-truth volumes. PADMix significantly outperforms previous literature on few-shot settings over the ShapeNet dataset and sets new benchmarks on the more challenging real-world Pix3D dataset.

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Builds on 33 of its 41 references that are in the map. First posted on arXiv 23 Dec 2021.

Builds on

33 of 41 references.

References · 33

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  3. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,880
  4. 2018Occupancy Networks: Learning 3D Reconstruction in Function Spacecited by 2,037
  5. 2016A Point Set Generation Network for 3D Object Reconstruction from a Single Imagecited by 972
  6. 2018Pixel2Mesh: Generating 3D Mesh Models from Single RGB Imagescited by 649
  7. 2016Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modelingcited by 558
  8. 2018Meta-Learning with Latent Embedding Optimizationcited by 476
  9. 2018Meta-Learning for Semi-Supervised Few-Shot Classificationcited by 456
  10. 2019A Baseline for Few-Shot Image Classificationcited by 298
  11. 2018Pix3D: Dataset and Methods for Single-Image 3D Shape Modelingcited by 218
  12. 2021Adaptive Prototype Learning and Allocation for Few-Shot Segmentationcited by 207
21 more
  1. 2020Prototype Mixture Models for Few-shot Semantic Segmentationcited by 190
  2. 2021FSCE: Few-Shot Object Detection via Contrastive Proposal Encodingcited by 168
  3. 2019Pix2Vox: Context-aware 3D Reconstruction from Single and Multi-view Imagescited by 164
  4. 2019Charting the Right Manifold: Manifold Mixup for Few-shot Learningcited by 135
  5. 2017Learning Efficient Point Cloud Generation for Dense 3D Object Reconstructioncited by 133
  6. 2020PointMixup: Augmentation for Point Cloudscited by 104
  7. 2019Pixel2Mesh++: Multi-View 3D Mesh Generation via Deformationcited by 103
  8. 2019Meta-Learning with Warped Gradient Descentcited by 98
  9. 2021Generalized Few-Shot Object Detection without Forgettingcited by 88
  10. 2020Pix2Vox++: Multi-scale Context-aware 3D Object Reconstruction from Single and Multiple Imagescited by 79
  11. 2018Learning Shape Priors for Single-View 3D Completion and Reconstructioncited by 68
  12. 2020Mask2CAD: 3D Shape Prediction by Learning to Segment and Retrievecited by 65
  13. 2019Associative Alignment for Few-shot Image Classificationcited by 58
  14. 2021Dense Relation Distillation with Context-aware Aggregation for Few-Shot Object Detectioncited by 53
  15. 2020CoReNet: Coherent 3D scene reconstruction from a single RGB imagecited by 45
  16. 2019TransMatch: A Transfer-Learning Scheme for Semi-Supervised Few-Shot Learningcited by 39
  17. 2018Learning View Priors for Single-view 3D Reconstructioncited by 35
  18. 2019Dense 3D Point Cloud Reconstruction Using a Deep Pyramid Networkcited by 33
  19. 2019Few-Shot Generalization for Single-Image 3D Reconstruction via Priorscited by 25
  20. 2021Fostering Generalization in Single-view 3D Reconstruction by Learning a Hierarchy of Local and Global Shape Priorscited by 9
  21. 2020Single-View 3D Object Reconstruction from Shape Priors in Memorycited by 3

Built on by

Cited by year

Citing papers per year: 2022: 2; 2023: 3.

5 papers.

  1. 2022Semi-Supervised Single-View 3D Reconstruction via Prototype Shape Priorscited by 14
  2. 2022Few-shot Single-view 3D Reconstruction with Memory Prior Contrastive Networkcited by 9
  3. 2023Robustmix: Improving Robustness by Regularizing the Frequency Bias of Deep Netscited by 3
  4. 2023Infinite Class Mixupcited by 2
  5. 20233DMiner: Discovering Shapes from Large-Scale Unannotated Image Datasetscited by 0