基于SMPL模态分解与嵌入融合的多模态步态识别

吴越,梁铮,高巍,杨茂达,赵培森,邓红霞,常媛媛

Multi-modal gait recognition based on SMPL model decomposition and embedding fusion

Yue WU,Zheng LIANG,Wei GAO,Maoda YANG,Peisen ZHAO,Hongxia DENG,Yuanyuan CHANG

表 1 轮廓分支ResNet类主干结构

Tab.1 ResNet-like backbone structure of silhouette branch

模块 模块结构 输出维度 Block0 $ \left[3\times 3, 64\right],\; {\rm{stride}}=1 $ 30×64×64×44 Block1 $ \left[\begin{array}{c}3\times 3, 64\\ 3\times 3, 64\end{array}\right],\;{\rm{stride}}=1 $ $ \left[3\times 1\times 1, 64\right], \;{\rm{stride}}=1 $ 30×128×64×44 Block2 $ \left[\begin{array}{c}3\times 3, 64\\ 3\times 3, 128\end{array}\right],\;{\rm{stride}}=2 $ $ \left[3\times 1\times 1, 128\right] , \;{\rm{stride}}=1 $ 30×128×32×22 Block3 $ \left[\begin{array}{c}3\times 3, 128\\ 3\times 3, 256\end{array}\right],\;{\rm{stride}}=2 $ $ \left[3\times 1\times 1, 256\right],\; {\rm{stride}}=1 $ 30×256×16×11 Block4 $ \left[\begin{array}{c}3\times 3, 256\\ 3\times \mathrm{3,256}\end{array}\right] ,\;{\rm{stride}}=1 $ $ \left[3\times 1\times 1, 256\right],\; {\rm{stride}}=1 $ 30×256×16×11