mlsanityData to 7 Oct 2026

Deep Deformable Models: Learning 3D Shape Abstractions with Part Consistency

arXiv 2309.01035 ↗First posted 2 Sep 2023cs.CV cs.AIcited by 5

Not on the map. Why?

Abstract (arXiv)

The task of shape abstraction with semantic part consistency is challenging due to the complex geometries of natural objects. Recent methods learn to represent an object shape using a set of simple primitives to fit the target. textcolorblackHowever, in these methods, the primitives used do not always correspond to real parts or lack geometric flexibility for semantic interpretation. In this paper, we investigate salient and efficient primitive descriptors for accurate shape abstractions, and propose textitDeep Deformable Models (DDMs). DDM employs global deformations and diffeomorphic local deformations. These properties enable DDM to abstract complex object shapes with significantly fewer primitives that offer broader geometry coverage and finer details. DDM is also capable of learning part-level semantic correspondences due to the differentiable and invertible properties of our primitive deformation. Moreover, DDM learning formulation is based on dynamic and kinematic modeling, which enables joint regularization of each sub-transformation during primitive fitting. Extensive experiments on textitShapeNet demonstrate that DDM outperforms the state-of-the-art in terms of reconstruction and part consistency by a notable margin.

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Builds on 33 of its 53 references that are in the map. First posted on arXiv 2 Sep 2023.

Builds on

33 of 53 references.

References · 33

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,384
  4. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,880
  5. 2019DeepSDF: Learning Continuous Signed Distance Functions for Shape Representationcited by 2,682
  6. 2018Occupancy Networks: Learning 3D Reconstruction in Function Spacecited by 2,037
  7. 2018Learning Implicit Fields for Generative Shape Modelingcited by 1,060
  8. 2016A Point Set Generation Network for 3D Object Reconstruction from a Single Imagecited by 972
  9. 20163D-R2N2: A Unified Approach for Single and Multi-view 3D Object Reconstructioncited by 587
  10. 2016Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modelingcited by 558
  11. 2018AtlasNet: A Papier-Mach'e Approach to Learning 3D Surface Generationcited by 539
  12. 2016Learning Shape Abstractions by Assembling Volumetric Primitivescited by 181
21 more
  1. 2019CvxNet: Learnable Convex Decompositioncited by 173
  2. 2019Superquadrics Revisited: Learning 3D Shape Parsing beyond Cuboidscited by 123
  3. 2020Deformed Implicit Field: Modeling 3D Shapes with Learned Dense Correspondencecited by 119
  4. 2018Unsupervised Learning for Fast Probabilistic Diffeomorphic Registrationcited by 116
  5. 2021Neural Parts: Learning Expressive 3D Shape Abstractions with Invertible Neural Networkscited by 101
  6. 2020Deep Implicit Templates for 3D Shape Representationcited by 86
  7. 2020DualSDF: Semantic Shape Manipulation using a Two-Level Representationcited by 82
  8. 20173D-PRNN: Generating Shape Primitives with Recurrent Neural Networkscited by 77
  9. 2020Learning Unsupervised Hierarchical Part Decomposition of 3D Objects from a Single RGB Imagecited by 72
  10. 2018Im2Struct: Recovering 3D Shape Structure from a Single RGB Imagecited by 55
  11. 2020KeypointNet: A Large-scale 3D Keypoint Dataset Aggregated from Numerous Human Annotationscited by 50
  12. 2022A Data-scalable Transformer for Medical Image Segmentation: Architecture, Model Efficiency, and Benchmarkcited by 45
  13. 2020Unsupervised Learning of Intrinsic Structural Representation Pointscited by 39
  14. 2023PartNeRF: Generating Part-Aware Editable 3D Shapes without 3D Supervisioncited by 36
  15. 2020Learning Implicit Functions for Topology-Varying Dense 3D Shape Correspondencecited by 27
  16. 2023Steering Prototypes with Prompt-tuning for Rehearsal-free Continual Learningcited by 22
  17. 2021Learning 3D Dense Correspondence via Canonical Point Autoencodercited by 21
  18. 2022TransFusion: Multi-view Divergent Fusion for Medical Image Segmentation with Transformerscited by 21
  19. 2022DeepRecon: Joint 2D Cardiac Segmentation and 3D Volume Reconstruction via A Structure-Specific Generative Methodcited by 17
  20. 2022Region Proposal Rectification Towards Robust Instance Segmentation of Biological Imagescited by 13
  21. 2023Dealing With Heterogeneous 3D MR Knee Images: A Federated Few-Shot Learning Method With Dual Knowledge Distillationcited by 11

Built on by

Cited by year

Citing papers per year: 2023: 1; 2024: 1; 2025: 3.

5 papers.

  1. 2023DeFormer: Integrating Transformers with Deformable Models for 3D Shape Abstraction from a Single Imagecited by 11
  2. 2025LUCAS: Layered Universal Codec Avatarscited by 7
  3. 2024Instantaneous Perception of Moving Objects in 3Dcited by 3
  4. 2025Snapmoji: Instant Generation of Animatable Dual-Stylized Avatarscited by 2
  5. 2025Large Sign Language Models: Toward 3D American Sign Language Translationcited by 1