SasMamba: A Lightweight Structure-Aware Stride State Space Model for 3D Human Pose Estimation
Hu Cui, Wenqiang Hua, Renjing Huang, Shurui Jia, Tessai Hayama
最近,基于国家空间模型(SSM)的Mamba架构因其线性复杂性和强大的全球建模能力而获得了3D人类姿势估计的关注。 然而,现有的基于SSM的方法通常应用手动设计的扫描操作,将检测到的2D姿势序列扁平成纯时间序列,无论是本地还是全球。 这种方法破坏了人类姿势的固有空间结构,并纠缠了时空特征,使得难以捕捉复杂的姿势依赖关系。 为了解决这些限制,我们提出了骷髅结构感知步幅SSM(SAS-SSM),它首先采用结构感知时空卷积来动态捕获关节之间的基本局部相互作用,然后应用基于步幅的扫描策略来构建多尺度的全球结构表示。 这可以实现本地和全局构图信息的灵活建模,同时保持线性计算复杂性。 基于SAS-SSM,我们的模型SasMamba实现了具有竞争力的3D姿势估计性能,与现有混合模型相比,参数要少得多。 源代码可在https://hucui2022.github.io/sasmamba_proj/上查阅。
Recently, the Mamba architecture based on State Space Models (SSMs) has gained attention in 3D human pose estimation due to its linear complexity and strong global modeling capability. However, existing SSM-based methods typically apply manually designed scan operations to flatten detected 2D pose sequences into purely temporal sequences, either locally or globally. This approach disrupts the inherent spatial structure of human poses and entangles spatial and temporal features, making it difficu...
