mlsanityData to 14 Oct 2026

Pixel3DMM: Versatile Screen-Space Priors for Single-Image 3D Face Reconstruction

arXiv 2505.00615 ↗First posted 1 May 2025cs.CV cs.AIcited by 20

Not on the map. Why?

Abstract (arXiv)

We address the 3D reconstruction of human faces from a single RGB image. To this end, we propose Pixel3DMM, a set of highly-generalized vision transformers which predict per-pixel geometric cues in order to constrain the optimization of a 3D morphable face model (3DMM). We exploit the latent features of the DINO foundation model, and introduce a tailored surface normal and uv-coordinate prediction head. We train our model by registering three high-quality 3D face datasets against the FLAME mesh topology, which results in a total of over 1,000 identities and 976K images. For 3D face reconstruction, we propose a FLAME fitting opitmization that solves for the 3DMM parameters from the uv-coordinate and normal estimates. To evaluate our method, we introduce a new benchmark for single-image face reconstruction, which features high diversity facial expressions, viewing angles, and ethnicities. Crucially, our benchmark is the first to evaluate both posed and neutral facial geometry. Ultimately, our method outperforms the most competitive baselines by over 15% in terms of geometric accuracy for posed facial expressions.

Key facts. Cited by 20 papers in the map (data through 14 Oct 2026). Builds on 41 of its 53 references that are in the map. First posted on arXiv 1 May 2025.

Builds on

41 of 53 references.

References · 41

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  3. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  4. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  5. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 6,013
  6. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,845
  7. 2022DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generationcited by 3,300
  8. 2021Vision Transformers for Dense Predictioncited by 1,803
  9. 2023DUSt3R: Geometric 3D Vision Made Easycited by 1,567
  10. 2020Face2Face: Real-time Face Capture and Reenactment of RGB Videoscited by 632
  11. 2017How far are we from solving the 2D & 3D Face Alignment problem? (and a dataset of 230,000 3D facial landmarks)cited by 525
  12. 2020Learning an Animatable Detailed 3D Face Model from In-The-Wild Imagescited by 477
29 more
  1. 2024Metric3Dv2: A Versatile Monocular Geometric Foundation Model for Zero-shot Metric Depth and Surface Normal Estimationcited by 395
  2. 2022EMOCA: Emotion Driven Monocular Face Capture and Animationcited by 255
  3. 2024Sapiens: Foundation for Human Vision Modelscited by 229
  4. 2020Towards Fast, Accurate and Stable 3D Dense Face Alignmentcited by 219
  5. 2021Fake It Till You Make It: Face analysis in the wild using synthetic data alonecited by 200
  6. 2021General Facial Representation Learning in a Visual-Linguistic Mannercited by 195
  7. 2023NeRSemble: Multi-view Radiance Field Reconstruction of Human Headscited by 194
  8. 2019Learning to Regress 3D Face Shape and Expression from an Image without 3D Supervisioncited by 176
  9. 2022Towards Metrical Reconstruction of Human Facescited by 166
  10. 2019PRNet: Self-Supervised Learning for Partial-to-Partial Registrationcited by 146
  11. 2018Look at Boundary: A Boundary-Aware Face Alignment Algorithmcited by 138
  12. 2022Next3D: Generative Neural Texture Rasterization for 3D-Aware Head Avatarscited by 124
  13. 2024SMIRK: 3D Facial Expressions through Analysis-by-Neural-Synthesiscited by 84
  14. 2024Fine-Tuning Image-Conditional Diffusion Models is Easier than You Thinkcited by 83
  15. 2019BlazeFace: Sub-millisecond Neural Face Detection on Mobile GPUscited by 74
  16. 2022Learning Neural Parametric Head Modelscited by 73
  17. 20223D face reconstruction with dense landmarkscited by 70
  18. 2024CAP4D: Creating Animatable 4D Portrait Avatars with Morphable Multi-View Diffusion Modelscited by 52
  19. 2022Visual Speech-Aware Perceptual 3D Facial Expression Reconstruction from Videoscited by 50
  20. 2023DiffusionAvatars: Deferred Diffusion for High-fidelity 3D Head Avatarscited by 42
  21. 2021FaceScape: 3D Facial Dataset and Benchmark for Single-View 3D Face Reconstructioncited by 36
  22. 2023MonoNPHM: Dynamic Head Reconstruction from Monocular Videoscited by 35
  23. 2022REALY: Rethinking the Evaluation of 3D Face Reconstructioncited by 33
  24. 2020Attention Mesh: High-fidelity Face Mesh Prediction in Real-timecited by 30
  25. 2024Video Depth without Video Modelscited by 29
  26. 2018Evaluation of Dense 3D Reconstruction from 2D Face Images in the Wildcited by 22
  27. 20243D Face Tracking from 2D Video through Iterative Dense UV to Image Flowcited by 20
  28. 2024Joker: Conditional 3D Head Synthesis with Extreme Facial Expressionscited by 18
  29. 2023Single-Shot Implicit Morphable Faces with Consistent Texture Parameterizationcited by 16

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 5; 2026: 15.

20 papers.

Sort:

  1. 2025FlexAvatar: Learning Complete 3D Head Avatars with Partial Supervisioncited by 10
  2. 2026UIKA: Fast Universal Head Avatar from Pose-Free Imagescited by 5
  3. 2026FastGHA: Generalized Few-Shot 3D Gaussian Head Avatars with Real-Time Animationcited by 4
  4. 2025FlexAvatar: Flexible Large Reconstruction Model for Animatable Gaussian Head Avatars with Detailed Deformationcited by 2
  5. 2026SemanticFace: Semantic Facial Action Estimation via Semantic Distillation in Interpretable Spacecited by 2
  6. 2026M3T: Discrete Multi-Modal Motion Tokens for Sign Language Productioncited by 2
  7. 2026Face Anything: 4D Face Reconstruction from Any Image Sequencecited by 2
  8. 2025GTAvatar: Bridging Gaussian Splatting and Texture Mapping for Relightable and Editable Gaussian Avatarscited by 1
  9. 2025Pix2NPHM: Learning to Regress NPHM Reconstructions From a Single Imagecited by 1
  10. 2026Skullptor: High Fidelity 3D Head Reconstruction in Seconds with Multi-View Normal Predictioncited by 1
  11. 2026Feed-forward Gaussian Registration for Head Avatar Creation and Editingcited by 1
  12. 2026AvatarPointillist: AutoRegressive 4D Gaussian Avatarizationcited by 1
8 more
  1. 2025Densemarks: Learning Canonical Embeddings for Human Heads Images via Point Trackscited by 0
  2. 2026GeoDiff4D: Geometry-Aware Diffusion for 4D Head Avatar Reconstructioncited by 0
  3. 2026Revisiting an Old Perspective Projection for Monocular 3D Morphable Models Regressioncited by 0
  4. 2026PerformRecast: Expression and Head Pose Disentanglement for Portrait Video Editingcited by 0
  5. 2026Profile-Specific 3DMM Regression from a Single Lateral Face Imagecited by 0
  6. 2026SuperFace: Preference-Aligned Facial Expression Estimation Beyond Pseudo Supervisioncited by 0
  7. 2026Loki: Representation over Architecture for Diffusion-Based Portrait Animationcited by 0
  8. 2026GeoFace: Consistent Multi-View Face Generation with Geometry-Constrained Diffusioncited by 0