SpatialAvatar-0: High-Quality 4D Head Avatar with Multi-Stage Reconstruction

• URL: http://arxiv.org/abs/2606.15659v1
• Date: Sun, 14 Jun 2026 07:55:57 GMT
• Title: SpatialAvatar-0: High-Quality 4D Head Avatar with Multi-Stage Reconstruction
• Abstract summary: High-quality 4D head avatars are central to telepresence, AR/VR, and digital-human interaction.<n>Generalizable feed-forward predictors and per-subject refiners are maturing in parallel.<n>We propose SpatialAvatar-0 on a shared FLAME-mesh-bound Gaussian representation.
• Score: 29.13831933508847
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: High-quality 4D head avatars from one or a few source portraits are central to telepresence, AR/VR, and digital-human interaction. 3D Gaussian Splatting (3DGS) has emerged as the dominant representation, with two complementary regimes (generalizable feed-forward predictors and per-subject refiners) maturing in parallel. However, existing feed-forward predictors are trained on a single dataset family with a hard-coded source count, inheriting the corresponding domain bias. Per-subject refiners require 300K--600K iterations and rely on adaptive densification that destroys upstream Gaussian layouts, preventing the two regimes from sharing a representation end-to-end. To bridge both regimes we propose SpatialAvatar-0 on a shared FLAME-mesh-bound Gaussian representation: a feed-forward generator with a parameter-free K-source mean-pool and a monocular-temporal to multi-view-spatial two-phase schedule that anchors against identity-prior collapse onto the smaller multi-view set. We further introduce a 10K-iter layout-preserving per-subject refinement loop that freezes the FLAME-binding and Gaussian count and replaces densification with a three-component anti-spike regularization. On VFHQ/HDTF cross-domain zero-shot we surpass the in-domain leader GAGAvatar by +1.5 dB PSNR despite never training on either test domain, and on the SplattingAvatar monocular benchmark we lead every reported metric, surpassing the 300K-iter GeoAvatar by +1.3 dB PSNR at up to 60x shorter per-subject schedule than common SOTA baselines. Website: https://spatialwalk.github.io/SpatialAvatar-0.

Related papers

• ATSplat: Compact Feed-forward 3D Gaussian Splatting with Adaptive Token Expansion [60.00320387969695]
  ATSplat is a feed-forward 3DGS framework that restores the adaptive allocation capability of 3DGS optimization.<n>From 12 input images at $512 times 960$ resolution, ATSplat completes reconstruction in less than a second using a single commercial GPU.
  arXiv  Detail & Related papers  (2026-07-22T17:54:44Z)
• Rendering-Aware Bayesian 3D Gaussian Splatting with Native Uncertainty and Adaptive Complexity Control [20.917423573702127]
  3DGS is a strong representation for real-time novel-view, but its standard training pipeline relies on point estimates and hand-tuned synthesiss.<n>We introduce a rendering-aware 3DGS framework that tracks Bayesian geometry with a Normal-Inverse-Wishart posterior over means and covariances.<n>In a fixed-budget 16-to-32 active-view task, native NIW acquisition improves PSNR by +0.453 dB and LPIPS by -0.0146 over a scoring-only 3-member standard-ensemble baseline.
  arXiv  Detail & Related papers  (2026-07-06T18:01:08Z)
• PointSplat: Compact Gaussian Splatting via Human-Centric Prediction [70.45838533838104]
  We propose PointSplat, a novel human-centric approach to producing 3D human representations from input views on the fly.<n>The proposed method first estimates a coarse geometric proxy and performs ray casting to prune redundant points.<n>It then employs a Point-Image Transformer to fuse appearance and geometry features, predicting Gaussian attributes in a single forward pass.
  arXiv  Detail & Related papers  (2026-06-30T17:59:05Z)
• ZipSplat: Fewer Gaussians, Better Splats [89.08529100444004]
  We propose ZipSplat, a token-based feed-forward model that decouples Gaussian placement from the pixel grid.<n>Cross- and self-attention refine these tokens, and a single trained model spans the quality-efficiency curve without retraining.<n>ZipSplat operates without ground-truth poses or intrinsics, yet sets a new state of the art on DL3DV and RealEstate10K with $sim6times$ fewer Gaussians than pixel-aligned methods.
  arXiv  Detail & Related papers  (2026-06-03T17:04:30Z)
• Scaling Parallel Sequence Models to Foundation-Scale Vision Encoders [123.58723804218151]
  Vision foundation models are bottlenecked by the quadratic cost of self-attention.<n>We present C-GSPN based on 2D spatial propagation propagation.
  arXiv  Detail & Related papers  (2026-05-30T14:29:43Z)
• FFAvatar: Few-Shot, Feed-Forward, and Generalizable Avatar Reconstruction [15.685572282342891]
  Avatar reconstruction has traditionally relied on per-subject optimization that requires hours of computation.<n>We introduce FFAvatar, a generalizable feed-forward framework that reconstructs high-quality, animatable 3D Gaussian head avatars in seconds.
  arXiv  Detail & Related papers  (2026-05-14T18:33:49Z)
• Two-View Accumulation as the Primary Training Lever for Hybrid-Capture Gaussian Splatting: A Variance-Decomposition View of When Gradient Surgery Helps [7.6889618752994595]
  Hybrid-capture novel view synthesis combines substantially different camera views.<n>Standard 3DGS is trained for 30K iterations with one rendered view per step.<n>We propose a variance-decomposition framework that predicts and explains this finding.
  arXiv  Detail & Related papers  (2026-04-29T17:45:51Z)
• iGaussian: Real-Time Camera Pose Estimation via Feed-Forward 3D Gaussian Splatting Inversion [62.09575122593993]
  iGaussian is a two-stage feed-forward framework that achieves real-time camera pose estimation through direct 3D Gaussian inversion.<n> Experimental results on the NeRF Synthetic, Mip-NeRF 360, and T&T+DB datasets demonstrate a significant performance improvement over previous methods.
  arXiv  Detail & Related papers  (2025-11-18T05:22:22Z)
• Linear Differential Vision Transformer: Learning Visual Contrasts via Pairwise Differentials [34.77694214755808]
  We introduce Visual-Contrast Attention (VCA), a drop-in replacement for Multi-Head Self-Attention (MHSA)<n>VCA injects an explicit notion of discrimination while reducing the theoretical complexity from O(N N C) to O(N n C) with n N.<n>The module adds fewer than 0.3M parameters to a DeiT-Tiny backbone, requires no extra FLOPs, and is wholly architecture-agnostic.
  arXiv  Detail & Related papers  (2025-11-02T07:04:12Z)
• SPFSplatV2: Efficient Self-Supervised Pose-Free 3D Gaussian Splatting from Sparse Views [18.814209805277503]
  SPFSplatV2, an efficient feed-forward framework for 3D Gaussian splatting from sparse multi-view images, is presented.<n>Method achieves state-of-the-art performance in both in-domain and out-of-domain novel view synthesis.
  arXiv  Detail & Related papers  (2025-09-21T21:37:56Z)
• InstantSplat: Sparse-view Gaussian Splatting in Seconds [91.77050739918037]
  We introduce InstantSplat, a novel approach for addressing sparse-view 3D scene reconstruction at lightning-fast speed.<n>InstantSplat employs a self-supervised framework that optimize 3D scene representation and camera poses.<n>It achieves an acceleration of over 30x in reconstruction and improves visual quality (SSIM) from 0.3755 to 0.7624 compared to traditional SfM with 3D-GS.
  arXiv  Detail & Related papers  (2024-03-29T17:29:58Z)
• MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Images [102.7646120414055]
  We introduce MVSplat, an efficient model that, given sparse multi-view images as input, predicts clean feed-forward 3D Gaussians. On the large-scale RealEstate10K and ACID benchmarks, MVSplat achieves state-of-the-art performance with the fastest feed-forward inference speed (22fps)
  arXiv  Detail & Related papers  (2024-03-21T17:59:58Z)
• Mega-NeRF: Scalable Construction of Large-Scale NeRFs for Virtual Fly-Throughs [54.41204057689033]
  We explore how to leverage neural fields (NeRFs) to build interactive 3D environments from large-scale visual captures spanning buildings or even multiple city blocks collected primarily from drone data. In contrast to the single object scenes against which NeRFs have been traditionally evaluated, this setting poses multiple challenges. We introduce a simple clustering algorithm that partitions training images (or rather pixels) into different NeRF submodules that can be trained in parallel.
  arXiv  Detail & Related papers  (2021-12-20T17:40:48Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.