mlsanityData to 7 Oct 2026

PercHead: Perceptual Head Model for Single-Image 3D Head Reconstruction & Editing

arXiv 2511.02777 ↗First posted 4 Nov 2025cs.CVcited by 6

Not on the map. Why?

Abstract (arXiv)

We present PercHead, a model for single-image 3D head reconstruction and disentangled 3D editing - two tasks that are inherently challenging due to ambiguity in plausible explanations for the same input. At the heart of our approach lies our novel perceptual loss based on DINOv2 and SAM 2.1. Unlike widely-adopted low-level losses like LPIPS, SSIM or L1, we rely on deep visual understanding of images and the resulting generalized supervision signals. We show that our new loss can be a drop-in replacement for standard losses and used to improve visual quality in high-frequency areas. We base our model architecture on Vision Transformers (ViTs), allowing us to decouple the 3D representation from the 2D input. We train our method on multi-view images for view-consistency and in-the-wild images for strong transferability to new environments. Our model achieves state-of-the-art performance in novel-view synthesis and, furthermore, exhibits exceptional robustness to extreme viewing angles. We also extend our base model to disentangled 3D editing by swapping the encoder and fine-tuning the network. A segmentation map controls geometry and either a text prompt or a reference image specifies appearance. We highlight the intuitive and powerful 3D editing capabilities through an interactive GUI. Project Page: https://antoniooroz.github.io/PercHead Video: https://www.youtube.com/watch?v=4hFybgTk4kE

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Builds on 48 of its 61 references that are in the map. First posted on arXiv 4 Nov 2025.

Builds on

48 of 61 references.

References · 48

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  3. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  4. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,339
  5. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,141
  6. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  7. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesiscited by 4,290
  8. 2024SAM 2: Segment Anything in Images and Videoscited by 2,784
  9. 2021Efficient Geometry-aware 3D Generative Adversarial Networkscited by 1,204
  10. 2016Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Networkcited by 1,199
  11. 2021Cascaded Diffusion Models for High Fidelity Image Generationcited by 1,155
  12. 2021Vector Quantized Diffusion Model for Text-to-Image Synthesiscited by 773
36 more
  1. 2024LGM: Large Multi-View Gaussian Model for High-Resolution 3D Content Creationcited by 730
  2. 2020pi-GAN: Periodic Implicit Generative Adversarial Networks for 3D-Aware Image Synthesiscited by 708
  3. 2021StyleNeRF: A Style-based 3D-Aware Generator for High-resolution Image Synthesiscited by 477
  4. 2020Learning an Animatable Detailed 3D Face Model from In-The-Wild Imagescited by 475
  5. 2021Pivotal Tuning for Latent-based Editing of Real Imagescited by 427
  6. 2022Rodin: A Generative Model for Sculpting 3D Digital Avatars Using Diffusioncited by 309
  7. 2021StyleSDF: High-Resolution 3D-Consistent Image and Geometry Generationcited by 289
  8. 2023Efficient Diffusion Training via Min-SNR Weighting Strategycited by 242
  9. 2021General Facial Representation Learning in a Visual-Linguistic Mannercited by 193
  10. 2023NeRSemble: Multi-view Radiance Field Reconstruction of Human Headscited by 190
  11. 2022VFHQ: A High-Quality Dataset and Benchmark for Video Face Super-Resolutioncited by 174
  12. 2019GANFIT: Generative Adversarial Network Fitting for High Fidelity 3D Face Reconstructioncited by 150
  13. 2022Realistic One-shot Mesh-based Head Avatarscited by 119
  14. 2024Arc2Face: A Foundation Model for ID-Consistent Human Facescited by 101
  15. 2022FaceVerse: a Fine-grained and Detail-controllable 3D Face Morphable Model from a Hybrid Datasetcited by 98
  16. 2024Generalizable and Animatable Gaussian Head Avatarcited by 95
  17. 2020AvatarMe: Realistically Renderable 3D Facial Reconstruction "in-the-wild"cited by 91
  18. 2022High-fidelity 3D GAN Inversion by Pseudo-multi-view Optimizationcited by 65
  19. 2024Portrait4D-v2: Pseudo Multi-View Data Creates Better 4D Head Synthesizercited by 64
  20. 2022ClipFace: Text-guided Editing of Textured 3D Morphable Modelscited by 59
  21. 2023Portrait4D: Learning One-Shot 4D Head Avatar Synthesis using Synthetic Datacited by 55
  22. 2024GGHead: Fast and Generalizable 3D Gaussian Headscited by 51
  23. 2021AvatarMe++: Facial Shape and BRDF Inference with Photorealistic Rendering-Aware GANscited by 49
  24. 2024NPGA: Neural Parametric Gaussian Avatarscited by 46
  25. 2023TriPlaneNet: An Encoder for EG3D Inversioncited by 41
  26. 2022FFHQ-UV: Normalized Facial UV-Texture Dataset for 3D Face Reconstructioncited by 39
  27. 2024RodinHD: High-Fidelity 3D Avatar Generation with Diffusion Modelscited by 36
  28. 2023HHAvatar: Gaussian Head Avatar with Dynamic Hairscited by 35
  29. 2024Cafca: High-quality Novel View Synthesis of Expressive Faces from Casual Few-shot Capturescited by 34
  30. 2021Fast-GANFIT: Generative Adversarial Network for High Fidelity 3D Face Reconstructioncited by 33
  31. 2023Control3Diff: Learning Controllable 3D Diffusion Models from Single-view Imagescited by 27
  32. 2022Self-Supervised Geometry-Aware Encoder for Style-Based 3D GAN Inversioncited by 24
  33. 2025Arc2Avatar: Generating Expressive 3D Avatars from a Single Image via ID Guidancecited by 16
  34. 2023Multi-task Image Restoration Guided By Robust DINO Featurescited by 14
  35. 2023PREIM3D: 3D Consistent Precise Image Attribute Editing from a Single Imagecited by 11
  36. 2025MTReD: 3D Reconstruction Dataset for Fly-over Videos of Maritime Domaincited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 2; 2026: 4.

6 papers.

  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. 2025AGORA: Adversarial Generation Of Real-time Animatable 3D Gaussian Head Avatarscited by 3
  4. 2026Any3DAvatar: Fast and High-Quality Full-Head 3D Avatar Reconstruction from Single Portrait Imagecited by 1
  5. 2026Multi-view Consistent 3D Gaussian Head Avatars 'without' Multi-view Generationcited by 1
  6. 2026Large-Scale High-Quality 3D Gaussian Head Reconstruction from Multi-View Capturescited by 0