Algorithm 1 Graph construction: Here, ${\delta }_{\mathrm{D}}$ and ${\delta }_{\mathrm{C}}$ are threshold values to split a node into four children.

1:Input: image $I=\left\{depth\right(\mathrm{D}),color(\mathrm{C}\left)\right\}$, init tree depth $t_{\mathrm{D}}=0$ 2:  3:Output: decomposed image with quadtree structure 4:  5:functionQuadtree($I,t_{\mathrm{D}}$) 6:  7: if ($t_{\mathrm{D}}>=t_{\max }$) then 8:  9:  return $node\left(I\right)$ 10:  11: else 12:  13:  ${\sigma }_{\mathrm{D}}=$ standard deviation of $I_{\mathrm{D}}$ 14:  15:  ${\delta }_{\mathrm{D}}=$ divider of $I_{\mathrm{D}}$ 16:  17:  ${\sigma }_{\mathrm{C}}=$ standard deviation of $I_{\mathrm{C}}$ 18:  19:  ${\delta }_{\mathrm{C}}=$ divider of $I_{\mathrm{C}}$ 20:  21:  if (${\sigma }_{\mathrm{D}}>{\delta }_{\mathrm{D}}\parallel {\sigma }_{\mathrm{C}}>{\delta }_{\mathrm{C}}$) then 22:  23:   $I_{\mathrm{sub}}$ = split I as four sub images 24:  25:   return QuadTree($I_{\mathrm{sub}},t_{\mathrm{D}}+1$) 26:  27:  end if 28:  29: end if 30:  31: end function