Editing animatable human avatars typically relies on sparse supervision, often a few edited keyframes, yet naively fitting a reconstructed avatar to these edits frequently causes identity leakage and pose-dependent temporal flicker. We argue that these failures are best understood as an ill-conditioned inversion: the available edited constraints do not sufficiently determine the latent directions responsible for the intended edit. We propose a conditioning-guided edited reconstruction framework that performs editing as a constrained inversion in a structured avatar latent space, restricting updates to a low-dimensional, part-specific edit subspace to prevent unintended identity changes. Crucially, we design the editing constraints during inversion by optimizing a conditioning objective derived from a local linearization of the full decoding-and-rendering pipeline, yielding an edit-subspace information matrix whose spectrum predicts stability and drives frame reweighting / keyframe activation. The resulting method operates on small subspace matrices and can be implemented efficiently (e.g., via Hessian-vector products), and improves stability under limited edited supervision.
Figures & tables
Figure 1: Task Overview. Given a source monocular video (1st row) and sparse edited keyframes (2nd row), our method produces temporally stable avatar edits that preserve identity, along with per-keyframe importance weights (3rd row). The edited avatar supports downstream applications such as novel view synthesis and animation (4th row).
Figure 2: Overview of our pipeline.
Method
LPIPS e↓
PSNR e↑
LPIPS b↓
WE e↓
WE b↓
3DGSAvatar
0.162
23.41
0.071
0.048
0.036
DGE
0.214
20.87
0.132
0.089
0.074
IDOL
0.187
22.05
0.098
0.063
0.051
Ours
0.128
25.62
0.043
0.031
0.024
Table 1: Results on our benchmark with sparse supervision (5 uniform keyframes). We report edit fidelity (masked), leakage (outside mask), and temporal stability (warping error). Subscripts e and b denote edited and background regions, respectively.
Figure 3: Edited keyframes and renderings at unseen time steps are shown. The first row contains an incorrect edit, while the second row exhibits inconsistent editing appearance, where Edit2 introduces an arm tattoo.
Figure 4: We set the supervision budget to K=3 over five candidate keyframes. As optimization progresses, the weight mass shifts away from an obviously incorrect edit (frame 5), and subsequently from a frame with inconsistent appearance (frame 4, hat mismatch). The final selection concentrates on frames 1–3, yielding a consistent keyframe set.
Method
LPIPSfull↓
WEedit↓
logdetS(w)↑
Full
0.082
0.031
4.21
A1
0.095
0.047
3.05
A2
0.111
0.058
4.07
A3
0.107
0.039
4.13
Table 2: Ablation study. See Sec. 5.3 for details.
High-fidelity 3D head avatars have reached near-photorealistic quality. While recent methods enable text-driven manipulation, they struggle to provide fine-grained localized control, often entangling features or lacking geometric consistency. Modifying geometry through natural language currently requires slow per-prompt optimization or compromises identity and rigging. We present ManifoldSplat, the first end-toend framework for language-guided semantic shape editing of animatable 3D Gaussian Splatting avatars reconstructed from monocular videos. By performing edits within the structured FLAME manifold rather than directly optimizing an unstructured Gaussian cloud, we strictly preserve identity and animation. We introduce DeltaRegion, a per-region disentangled Conditional Variational Autoencoder (CVAE) delivering feedforward shape deltas, alongside a refining stage to recover view-consistent details. ManifoldSplat reconstructs and edits an avatar in ~90 seconds on a consumer GPU, rendering at ~800 FPS. Extensive evaluations demonstrate our approach sets a new state-of-the-art in localized prompt alignment, geometric coherence, and identity preservation. Project page and code: https://a-canela.github.io/manifoldsplat/
Antonio Canela, Jordi Sànchez-Riera
Universitat Politècnica de Catalunya · Institut de Robòtica i Informàtica Industrial, CSIC-UPC
We propose FlowAnchor, a training-free framework for stable and efficient inversion-free, flow-based video editing. Inversion-free editing methods have recently shown impressive efficiency and structure preservation in images by directly steering the sampling trajectory with an editing signal. However, extending this paradigm to videos remains challenging, often failing in multi-object scenes or with increased frame counts. We identify the root cause as the instability of the editing signal in high-dimensional video latent spaces, which arises from imprecise spatial localization and length-induced magnitude attenuation. To overcome this challenge, FlowAnchor explicitly anchors both where to edit and how strongly to edit. It introduces Spatial-aware Attention Refinement, which enforces consistent alignment between textual guidance and spatial regions, and Adaptive Magnitude Modulation, which adaptively preserves sufficient editing strength. Together, these mechanisms stabilize the editing signal and guide the flow-based evolution toward the desired target distribution. Extensive experiments demonstrate that FlowAnchor achieves more faithful, temporally coherent, and computationally efficient video editing across challenging multi-object and fast-motion scenarios. The project page is available at https://cuc-mipg.github.io/FlowAnchor.github.io/.
Ze Chen, Lan Chen, Yuanhang Li +1
MIPG, Communication University of China Beijing, China
Avatar reconstruction has traditionally relied on per-subject optimization that requires hours of computation or on expensive preprocessing that limits scalability. We introduce FFAvatar, a generalizable feed-forward framework that reconstructs high-quality, animatable 3D Gaussian head avatars from few-shot unposed portrait images in seconds. FFAvatar fuses information from multiple source images into a unified canonical Gaussian representation through Multi-View Query-Former, which is animated via FLAME parameters predicted end-to-end directly from pixels, eliminating the overhead of offline FLAME extraction. We further propose a three-stage training curriculum that achieves both broad generalization and high-fidelity reconstruction: (i) scalable pretraining on extensive monocular video data with over 1M identities to learn strong generalizable priors; (ii) multi-view fine-tuning on a small but high-quality dataset of 360-degree captures to enhance geometric fidelity and extreme-view awareness; and (iii) optional personalization that adapts to specific identities for maximum fidelity within 500 optimization steps. Extensive experiments demonstrate that FFAvatar sets a new standard for identity preservation, geometric consistency, and animation fidelity. On the NeRSemble benchmark, it outperforms the state-of-the-art LAM by a substantial 5.5 PSNR gain. Furthermore, FFAvatar enables real-time deployment, reconstructing avatars in 2 seconds without personalization and 10 seconds with personalization, while supporting 49 FPS animation on a single NVIDIA A100 GPU.
Thuan Hoang Nguyen, Jiahao Luo, Yinyu Nie +3
Snap Inc. · MBZUAI · University of California, Santa Cruz