Neural Motion Blending Across Arbitrary Character Topologies

2026-07-11

AI Summary

Existing character animation methods struggle to achieve effective motion blending across heterogeneous skeletal topologies. This work proposes a novel framework based on a semantic encoder and a conditional diffusion decoder: it first extracts frame-level latent representations of motion using a semantic encoder, performs interpolation in the latent space, and then decodes the result into character-specific, natural motion sequences via a diffusion model. The approach is the first to enable smooth motion blending between arbitrary skeletal structures, overcoming the longstanding reliance of conventional methods on identical or structurally similar skeletons. Experiments on the Truebones Zoo dataset demonstrate that the method consistently generates fluid and semantically coherent animations across diverse cross-topology scenarios.

Abstract

Motion blending in character animation enables the synthesis of new motions by interpolating between existing examples. Current methods are typically restricted to fixed skeleton topologies, requiring identical or near-identical skeletal structures across characters. We present a novel framework for motion blending across heterogeneous skeletons. The proposed architecture combines a semantic encoder, which extracts per-frame latent representations of the motion state, with a diffusion-based decoder, which reconstructs character-specific motion conditioned on this latent code. At inference, blended motions are obtained by interpolating the latent representations of two input motions. We train and evaluate the method on the Truebones Zoo dataset using motions defined on both same and distinct skeleton topologies, demonstrating the ability to achieve smooth and plausible blending in a variety of scenarios.

Problem

• motion blending
• character animation
• skeleton topology
• heterogeneous skeletons
• neural motion

Innovation

• neural motion blending
• heterogeneous skeletons
• semantic encoder
• diffusion-based decoder
• latent interpolation