Paper 2602.03796
3D-Aware Implicit Motion Control for View-Adaptive Human Video Generation
- Published
- Feb 2026
- Research lab
- Independent
- Citations
- 3
- GitHub
- Not linked
01 In brief
Summary
The paper introduces 3DiMo, a framework for 3D-aware human motion control in video generation.
It addresses limitations of existing methods that use 2D poses (viewpoint-bound) or explicit 3D models like SMPL (inaccurate, overriding generator priors).
3DiMo jointly trains a Transformer-based motion encoder with a pretrained DiT video generator to distill driving frames into compact, view-agnostic motion tokens, injected via cross-attention.
Training uses a view-rich dataset (internet, UE-rendered, in-house multi-view) with same-view reconstruction and cross-view reproduction objectives, plus auxiliary geometric supervision (SMPL/MANO) that is annealed to zero.
This enables text-guided camera control as a byproduct.
Experiments show 3DiMo outperforms baselines (AnimateAnyone, MimicMotion, MTVCrafter, Uni3C) on LPIPS, FID, FVD, and user study metrics, with ablations confirming the importance of implicit representations, multi-stage training, cross-attention, geometric supervision, and dual-scale encoders.
Applications include novel view synthesis, video stabilization, and automatic motion alignment.
02 From the paper
Abstract
Existing methods for human motion control in video generation typically rely on either 2D poses or explicit 3D parametric models (e.g., SMPL) as control signals. However, 2D poses rigidly bind motion to the driving viewpoint, precluding novel-view synthesis. Explicit 3D models, though structurally informative, suffer from inherent inaccuracies (e.g., depth ambiguity and inaccurate dynamics) which, when used as a strong constraint, override the powerful intrinsic 3D awareness of large-scale video generators. In this work, we revisit motion control from a 3D-aware perspective, advocating for an implicit, view-agnostic motion representation that naturally aligns with the generator's spatial priors rather than depending on externally reconstructed constraints. We introduce 3DiMo, which jointly trains a motion encoder with a pretrained video generator to distill driving frames into compact, view-agnostic motion tokens, injected semantically via cross-attention. To foster 3D awareness, we train with view-rich supervision (i.e., single-view, multi-view, and moving-camera videos), forcing motion consistency across diverse viewpoints. Additionally, we use auxiliary geometric supervision that leverages SMPL only for early initialization and is annealed to zero, enabling the model to transition from external 3D guidance to learning genuine 3D spatial motion understanding from the data and the generator's priors. Experiments confirm that 3DiMo faithfully reproduces driving motions with flexible, text-driven camera control, significantly surpassing existing methods in both motion fidelity and visual quality.