面向解耦文本驱动运动生成的语义先验分层增强
Hierarchical Enhancement of Semantic Priors for Disentangled Text-Driven Motion Generation
把潜空间结构化为多个语义一致子流形（AG-VAE）的思路，可直接迁移到视觉扩散模型的潜空间解耦与可控生成；分层跨模态注意力与动态记忆融合也可用于文本到图像/视频的条件注入设计。
原文依据 HESP introduces an Adaptive Gaussian Variational Autoencoder (AG-VAE) that structures the latent motion manifold into multiple semantically coherent submanifolds
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论文本到运动生成旨在从自然语言描述中合成真实且语义对齐的三维人体运动。现有基于扩散的方法往往依赖各向同性的潜变量先验与浅层的跨模态监督，从而导致语义纠缠、可控性受限以及可解释性较差。我们提出 HESP，一个统一的扩散框架，通过分层增强语义先验来实现解耦的文本驱动运动生成。HESP 的核心是引入自适应高斯变分自编码器（AG-VAE），将潜空间运动流形结构化为多个语义一致的子流形，从而获得可解释且可控的运动表征。为进一步弥合语言语义与运动学语义之间的鸿沟，我们设计了动态跨模态记忆（DCMM）模块以实现自适应语义融合，并设计分层跨模态注意力（HCA）机制以捕捉多层次的文本-运动对应关系。在 HumanML3D 与 KIT-ML 上的大量实验表明，HESP 持续优于 SALAD、MoMask 与 MDM 等最优基线，在取得提升的同时保持了更高的多样性与物理合理性。此外，HESP 的结构化潜空间提供了可解释的聚类，揭示出不同运动类别之间清晰的语义边界。
英文原文摘要 1341 字符
Text-to-motion generation aims to synthesize realistic and semantically aligned 3D human motions from natural language descriptions. Existing diffusion-based methods often rely on isotropic latent priors and shallow cross-modal supervision, which lead to semantic entanglement, limited controllability, and poor interpretability. We propose HESP, a unified diffusion framework that hierarchically enhances semantic priors for disentangled text-driven motion generation. At its core, HESP introduces an Adaptive Gaussian Variational Autoencoder (AG-VAE) that structures the latent motion manifold into multiple semantically coherent submanifolds, enabling interpretable and controllable motion representations. To further bridge linguistic and kinematic semantics, we design a Dynamic Cross-Modal Memory (DCMM) module for adaptive semantic fusion and a Hierarchical Cross-Modal Attention (HCA) mechanism to capture multi-level text-motion correspondences. Extensive experiments on HumanML3D and KIT-ML demonstrate that HESP consistently outperforms state-of-the-art baselines such as SALAD, MoMask, and MDM, achieving improvement while maintaining higher diversity and physical plausibility. Moreover, the structured latent space of HESP provides interpretable clusters that reveal clear semantic boundaries among different motion categories.