Nonnegative Tensor-Based Linear Dynamical Systems for Recognizing Human Action from 3D Skeletons

<table class="fixed-width table-group" id="tab6"><tr><td><table class="table"><colgroup><col style="width:12.98em"/><col style="width:23.57em"/><col style="width:23.57em"/></colgroup><tr><td class="thead-hr" colspan="3"><hr/></td></tr><tr class="thead"><td class="align_left">Dataset</td><td class="align_center">Tensor Representation + infinite Grassmann manifold</td><td class="align_center">Our approach<br/>(Nonnegative Tensor Representation + Infinite Grassmann Manifold)</td></tr><tr><td class="thead-hr" colspan="3"><hr/></td></tr><tr><td class="align_left">MSR-Action3D</td><td class="align_center">96.17</td><td class="align_center">96.97</td></tr><tr><td class="align_left" colspan="3"><hr/></td></tr><tr><td class="align_left">UTKinect-Action</td><td class="align_center">97.83</td><td class="align_center">98.23</td></tr><tr><td class="align_left" colspan="3"><hr/></td></tr><tr><td class="align_left">G3D-Gaming</td><td class="align_center">91.6</td><td class="align_center">92.56</td></tr><tr class="table-tr"><td colspan="3"><hr class="tbody-hr"/></td></tr></table></td></tr></table>

<div>Discussion about nonnegative tensor representation.</div>

Mathematical Problems in Engineering

tab6

Table 6

Table 6: Nonnegative Tensor-Based Linear Dynamical Systems for Recognizing Human Action from 3D Skeletons 