Algorithm 1. Identifying the bounding boxes of each skeleton.

1.  Input:  $skeletons=\{bod{y}_i\{j:(x_{joint\_j},y_{joint\_j}),j=[0,17]\},i=[0,n-1]\}$      $boxes\_object=\{bo{x}_i(x_1,y_1,x_2,y_2),i=[0,m-1]\}$      n: number of skeletons      m: number of bounding boxes

2.  Output: $boxes\_skeleton=\{bo{x}_i(x_1,y_1,x_2,y_2),i=[0,n-1]\}$      $boxes\_all=\{bo{x}_j(x_1,y_1,x_2,y_2),j=[0,k-1]\}$      k: number of persons

3.  for $body$ in $skeletons$:

4.   $x_1$ = min$\left(x_{body}\right)$

5.   $y_1$ = min$\left(y_{body}\right)$

6.   $x_2$ = max$\left(x_{body}\right)$

7.   $y_2$ = max$\left(y_{body}\right)$

8.   $boxes\_skeleton$.append$(x_1,y_1,x_2,y_2)$

9.   $boxes\_all=boxes\_skeleton$

10.  $status\_skeletons$ = [True for (i, $item$) in enumerate($skeletons$)]

11.  for ($index\_box$, $box$) in enumerate($boxes\_object$):

12:         $index\_max$ = -1

13.         $number\_joints$ = 0

14.         for ($index\_body$, $body$) in enumerate($skeletons$):

15.     if not(status_skeletons[$index\_body$]):

16.        continue

17.      if $number\_joints$ < count_points($box$, $body$):

18.        $number\_joints$ = count_points($box$, $body$)

19.        $index\_max=index\_body$

20.   if $index\_max$ == -1:

21.         $boxes\_all$.append($box$)

22.   else:

23.         $status\_skeletons[index\_max]$ = False

24.         $boxes\_skeleton[index\_max]=box$

25.         $boxes\_all[index\_max]=box$

26.   return $boxes\_skeleton,boxes\_all$