mlsanityData to 7 Oct 2026

3D Generative Model Latent Disentanglement via Local Eigenprojection

arXiv 2302.12798 ↗First posted 24 Feb 2023cs.CV cs.GR cs.LGcited by 6

Not on the map. Why?

Abstract (arXiv)

Designing realistic digital humans is extremely complex. Most data-driven generative models used to simplify the creation of their underlying geometric shape do not offer control over the generation of local shape attributes. In this paper, we overcome this limitation by introducing a novel loss function grounded in spectral geometry and applicable to different neural-network-based generative models of 3D head and body meshes. Encouraging the latent variables of mesh variational autoencoders (VAEs) or generative adversarial networks (GANs) to follow the local eigenprojections of identity attributes, we improve latent disentanglement and properly decouple the attribute creation. Experimental results show that our local eigenprojection disentangled (LED) models not only offer improved disentanglement with respect to the state-of-the-art, but also maintain good generation capabilities with training times comparable to the vanilla implementations of the models.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Builds on 57 of its 92 references that are in the map. First posted on arXiv 24 Feb 2023.

Builds on

57 of 92 references.

References · 57

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2016Convolutional Neural Networks on Graphs with Fast Localized Spectral Filteringcited by 2,085
  3. 2021Alias-Free Generative Adversarial Networkscited by 1,181
  4. 2013Spectral Networks and Locally Connected Networks on Graphscited by 1,097
  5. 2016Least Squares Generative Adversarial Networkscited by 1,000
  6. 2019MaskGAN: Towards Diverse and Interactive Facial Image Manipulationcited by 624
  7. 2018Disentangling by Factorisingcited by 614
  8. 2020GANSpace: Discovering Interpretable GAN Controlscited by 612
  9. 2019PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flowscited by 487
  10. 2020InterFaceGAN: Interpreting the Disentangled Face Representation Learned by GANscited by 393
  11. 2017Learning from Synthetic Humanscited by 358
  12. 20163D Menagerie: Modeling the 3D shape and pose of animalscited by 310
38 more
  1. 2020Unsupervised Discovery of Interpretable Directions in the GAN Latent Spacecited by 288
  2. 20193D Morphable Face Models -- Past, Present and Futurecited by 269
  3. 2018Generating 3D faces using Convolutional Mesh Autoencoderscited by 267
  4. 2017AttGAN: Facial Attribute Editing by Only Changing What You Wantcited by 241
  5. 2017Variational Inference of Disentangled Latent Concepts from Unlabeled Observationscited by 240
  6. 2020STAR: Sparse Trained Articulated Human Body Regressorcited by 194
  7. 2021EditGAN: High-Precision Semantic Image Editingcited by 156
  8. 2015Deep Convolutional Inverse Graphics Networkcited by 121
  9. 2020Learning Formation of Physically-Based Face Attributescited by 94
  10. 2021GAN-Control: Explicitly Controllable GANscited by 91
  11. 2019Neural 3D Morphable Models: Spiral Convolutional Networks for 3D Shape Representation Learning and Generationcited by 89
  12. 2019SpiralNet++: A Fast and Highly Efficient Mesh Convolution Operatorcited by 89
  13. 2021Multimodal Conditional Image Synthesis with Product-of-Experts GANscited by 75
  14. 2022ImFace: A Nonlinear 3D Morphable Face Model with Implicit Neural Representationscited by 64
  15. 2019Towards a complete 3D morphable model of the human headcited by 63
  16. 2018Structured Disentangled Representationscited by 62
  17. 2020Unsupervised Shape and Pose Disentanglement for 3D Meshescited by 61
  18. 2020Guided Variational Autoencoder for Disentanglement Learningcited by 57
  19. 2020LIMP: Learning Latent Shape Representations with Metric Preservation Priorscited by 54
  20. 2020Fully Convolutional Mesh Autoencoder using Efficient Spatially Varying Kernelscited by 53
  21. 2017Deformable Shape Completion with Graph Convolutional Autoencoderscited by 50
  22. 2019Combining 3D Morphable Models: A Large scale Face-and-Head Modelcited by 49
  23. 2019Disentangled Representation Learning for 3D Face Shapecited by 48
  24. 2021Self-Supervised Learning Disentangled Group Representation as Featurecited by 47
  25. 2021ARAPReg: An As-Rigid-As Possible Regularization Loss for Learning Deformable Shape Generatorscited by 40
  26. 2019MeshGAN: Non-linear 3D Morphable Models of Facescited by 39
  27. 2020Disentangled Image Generation Through Structured Noise Injectioncited by 39
  28. 2019Synthesizing Coupled 3D Face Modalities by Trunk-Branch Generative Adversarial Networkscited by 34
  29. 2020Codimensional Incremental Potential Contactcited by 33
  30. 2020DSG-Net: Learning Disentangled Structure and Geometry for 3D Shape Generationcited by 31
  31. 2019Geometric Disentanglement for Generative Latent Shape Modelscited by 30
  32. 20193DFaceGAN: Adversarial Nets for 3D Face Representation, Generation, and Translationcited by 27
  33. 2019Backpropagation-Friendly Eigendecompositioncited by 27
  34. 2021EditVAE: Unsupervised Part-Aware Controllable 3D Point Cloud Shape Generationcited by 21
  35. 2021Intrinsic-Extrinsic Preserved GANs for Unsupervised 3D Pose Transfercited by 20
  36. 2020Learning Disentangled Representations with Latent Variation Predictabilitycited by 19
  37. 2021Sparse to Dense Dynamic 3D Facial Expression Generationcited by 18
  38. 2021Local Disentanglement in Variational Auto-Encoders Using Jacobian L1 Regularizationcited by 17

50 most cited of 57 All 57 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

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

6 papers.

  1. 2024ShapeFusion: A 3D diffusion model for localized shape editingcited by 6
  2. 2024Locally Adaptive Neural 3D Morphable Modelscited by 5
  3. 2025ImHead: A Large-scale Implicit Morphable Model for Localized Head Modelingcited by 4
  4. 2026Parallelised Differentiable Straightest Geodesics for 3D Meshescited by 2
  5. 2023Latent Disentanglement in Mesh Variational Autoencoders Improves the Diagnosis of Craniofacial Syndromes and Aids Surgical Planningcited by 0
  6. 2026ARAPDiffusion: ARAP Regularization for Diffusion-Based Deformable Shape Space Learningcited by 0