Make-It-Poseable: Feed-forward Latent Posing Model for 3D Humanoid Character Animation

2025-12-18

AI Summary

Existing 3D character posing methods suffer from inaccurate skinning weights, topology sensitivity, and pose inconsistency, limiting generalization and robustness. This paper introduces a novel latent-space pose transformation paradigm that replaces conventional vertex-based deformation with feed-forward mapping in a learned latent space. Our key contributions are: (1) a skeletal-motion-aware Transformer architecture operating directly in latent space; (2) a dense, disentangled latent pose representation enabling fine-grained control; and (3) a joint latent-space supervision scheme coupled with an adaptive geometric completion module, explicitly accommodating topology changes. Experiments demonstrate substantial improvements in pose fidelity and cross-topology generalization—enabling high-fidelity part replacement, localized detail editing, and robust performance across diverse mesh topologies and articulations.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Learning with Manifolds
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Posing 3D characters is a fundamental task in computer graphics and vision. However, existing methods like auto-rigging and pose-conditioned generation often struggle with challenges such as inaccurate skinning weight prediction, topological imperfections, and poor pose conformance, limiting their robustness and generalizability. To overcome these limitations, we introduce Make-It-Poseable, a novel feed-forward framework that reformulates character posing as a latent-space transformation problem. Instead of deforming mesh vertices as in traditional pipelines, our method reconstructs the character in new poses by directly manipulating its latent representation. At the core of our method is a latent posing transformer that manipulates shape tokens based on skeletal motion. This process is facilitated by a dense pose representation for precise control. To ensure high-fidelity geometry and accommodate topological changes, we also introduce a latent-space supervision strategy and an adaptive completion module. Our method demonstrates superior performance in posing quality. It also naturally extends to 3D editing applications like part replacement and refinement.

Problem

• Improves 3D character posing via latent-space transformation
• Addresses inaccurate skinning and poor pose conformance issues
• Enables high-fidelity animation and 3D editing applications

Innovation

• Latent-space transformation for character posing
• Latent posing transformer with skeletal motion
• Dense pose representation and adaptive completion