Face and shape: research map
2,246 papers on Face and shape in 3D vision, from ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR (2016–2026), grouped into 4 clusters and 17 approaches. The busiest year so far is 2026.
What it is
Modelling the 3D shape of objects, faces and people: reconstructing them from images, generating new shapes, building avatars that can be animated, and the representations used (meshes, implicit surfaces, morphable models).
Terms you will see on this page:
- Implicit surface
- a surface defined as where a learned function equals zero.
- 3D morphable model (3DMM)
- a parametric model of face shape and expression.
- Avatar
- an animatable 3D model of a person.
Start reading with the most cited papers here:
- DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation (CVPR 2019 · 3,528 citations)
- Occupancy Networks: Learning 3D Reconstruction in Function Space (CVPR 2019 · 2,830 citations)
Within 3D vision, its share shrank from 20.4% in 2023–24 to 13.2% in 2025–26 (363 → 366 papers at ICML, NeurIPS, ICLR, CVPR and AAAI, the venues with data for all four years).
Explore Face and shape in the interactive map
Working on something in this topic? Describe your idea in scime atlas to see which approach it falls under, the closest papers by meaning and how crowded the spot has become.
What changed recently
Approaches whose share of the topic moved most between 2023–24 and 2025–26 (ICML, NeurIPS, ICLR, CVPR and AAAI):
- Textured mesh generation rising: 6% → 16% of the topic (21 → 60 papers)
- Implicit neural representation design rising: 4% → 7% of the topic (13 → 26 papers)
- 3D shape generative modeling rising: 5% → 7% of the topic (17 → 25 papers)
- Face reflectance and relightable rendering falling: 7% → 2% of the topic (25 → 9 papers)
- 3D morphable face model regression falling: 5% → 1% of the topic (19 → 5 papers)
Papers from 2025–2026 that are already cited most:
- Dora: Sampling and Benchmarking for 3D Shape Variational Auto-Encoders (CVPR 2025 · 20 citations)
- Adaptive Wavelet-Positional Encoding for High-Frequency Information Learning in Implicit Neural Representation (AAAI 2025 · 6 citations)
Approaches and key papers
3D shape modelling · 764 papers
Keywords: 3d shape, shapes, object
Approaches in this cluster:
- Primitive-based shape abstraction (187 papers)
Represents shapes with learned primitives, parts and implicit grids for retrieval and generation. - Single-view 3D reconstruction (190 papers)
Reconstructs 3D shape from one or few images with hierarchical priors and generalization to unseen classes. - Articulated and non-rigid shape learning (134 papers)
Learns deformable and articulated shapes and dense correspondence from casual videos and image collections. - 3D shape generative modeling (120 papers)
Generates 3D shapes with deep generative models, latent grids, primitives and text conditioning. - Optimization-based shape matching (133 papers)
Computes shape correspondence using spectral, combinatorial and registration-based solvers.
Most cited and most cited since 2024:
- Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling (NeurIPS 2016 · 1,544 citations)
- Learning Implicit Fields for Generative Shape Modeling (CVPR 2019 · 1,503 citations)
- Diffusion 3D Features (Diff3F): Decorating Untextured Shapes with Distilled Semantic Features (CVPR 2024 · 29 citations)
- Learning the 3D Fauna of the Web (CVPR 2024 · 23 citations)
Surfaces and meshes · 641 papers
Keywords: surface, mesh, implicit
Approaches in this cluster:
- Neural implicit surface reconstruction (244 papers)
Reconstructs surfaces from multi-view images with signed or unsigned distance fields and volume rendering. - Textured mesh generation (160 papers)
Generates textured and PBR 3D meshes with generative models and sparse high-resolution representations. - Learned mesh and geometric representations (105 papers)
Designs neural networks that produce meshes and fields with topology, analytic or vector-field representations. - Implicit neural representation design (75 papers)
Studies architectures, spectral bias and Fourier features for implicit neural representations. - Pixel-aligned implicit human and hand reconstruction (57 papers)
Reconstructs humans and hand-held objects with pixel-aligned implicit functions and explicit contact.
Most cited and most cited since 2024:
- DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation (CVPR 2019 · 3,528 citations)
- Occupancy Networks: Learning 3D Reconstruction in Function Space (CVPR 2019 · 2,830 citations)
- MeshGPT: Generating Triangle Meshes with Decoder-Only Transformers (CVPR 2024 · 113 citations)
- Intraoperative 2D/3D Image Registration via Differentiable X-ray Rendering (CVPR 2024 · 47 citations)
Human and head avatars · 472 papers
Keywords: avatar, human, body
Approaches in this cluster:
- Gaussian and implicit head avatars (184 papers)
Builds animatable head avatars from video using morphable, Gaussian blendshape or implicit representations. - Full-body avatar reconstruction (145 papers)
Reconstructs digital humans and articulated objects from monocular or single-view inputs with diffusion or disentangled modeling. - Clothed human modeling (105 papers)
Models and reconstructs clothing and clothed bodies using implicit surfaces and template decomposition. - Generative textures for 3D humans (38 papers)
Synthesizes editable textures and disentangled geometry-appearance using adversarial and neural rasterization.
Most cited and most cited since 2024:
- Neural Body: Implicit Neural Representations With Structured Latent Codes for Novel View Synthesis of Dynamic Humans (CVPR 2021 · 633 citations)
- Total Capture: A 3D Deformation Model for Tracking Faces, Hands, and Bodies (CVPR 2018 · 548 citations)
- GaussianAvatars: Photorealistic Head Avatars with Rigged 3D Gaussians (CVPR 2024 · 241 citations)
- GaussianAvatar: Towards Realistic Human Avatar Modeling from a Single Video via Animatable 3D Gaussians (CVPR 2024 · 171 citations)
3D faces · 369 papers
Keywords: 3d face, face reconstruction, faces
Approaches in this cluster:
- 3D morphable face model regression (166 papers)
Regresses morphable face parameters from images using deep networks, graph convolutions and implicit models. - Face reflectance and relightable rendering (145 papers)
Recovers facial reflectance and lighting with inverse and neural rendering. - Speech-driven 3D facial animation (58 papers)
Synthesizes expressive 3D facial motion from audio using discrete priors and probabilistic generation.
Most cited and most cited since 2024:
- RetinaFace: Single-Shot Multi-Level Face Localisation in the Wild (CVPR 2020 · 1,623 citations)
- Face Alignment Across Large Poses: A 3D Solution (CVPR 2016 · 1,211 citations)
- SyncTalk: The Devil is in the Synchronization for Talking Head Synthesis (CVPR 2024 · 76 citations)
- 3D Facial Expressions through Analysis-by-Neural-Synthesis (CVPR 2024 · 48 citations)
Related topics in 3D vision
- Gaussian splatting (1,340)
- Human pose estimation (2,063)
- NeRF and radiance fields (733)
- 3D generation (2,448)
- Depth and stereo (2,574)
