Algorithm 1. Grid Based Greedy Gait Segmentation Algorithm *
1	Input: $\mathsf{\epsilon }$, $\mathcal{C}$, $N$// $\mathsf{\epsilon }$ is the termination criteria for the iteration
2	Initialize: $c_0=\widehat{g}$, $b_0=0$, $i=0$, $h_0\left(k\right)=b_0+\left(k-1\right)c_0$ for $k=0:N_c-1$
3	Do:// equally spaced grid optimization by iteration
4	$i=i+1$, $N_c=\mathrm{int}(N/c_{i-1}$),
5	For $k=0:N_c-1$ Do:
6	$h_{i-1}\left(k\right)=\mathrm{int}(b_{i-1}+\left(k-1\right)c_{i-1})$,
7	$\widehat{\varsigma }\left(k\right)$ = GreedySearch1($h_{i-1}\left(k\right)$, $\mathcal{C}$),
8	End For
9	$c_i={\sum }_{k=1}^{N_c-1}\left(\widehat{\varsigma }\left(k\right)-\widehat{\varsigma }\left(k-1\right)\right)/(N_c-1)$,
10	$b_i=b_{i-1}+{\sum }_{k=0}^{N_c-1}\left(\widehat{\varsigma }\left(k\right)-h_{i-1}\left(k\right)\right)/N_c$,
11	While: $\max \left(b_i-b_{i-1},c_i-c_{i-1}\right)>\mathsf{\epsilon }$
12	$N_c=\mathrm{int}(N/c_i$),
13	For $k=0:N_c-1$ Do: //locally optimized without the equally spaced constraint
14	$h_i\left(k\right)=\mathrm{int}(b_i+\left(k-1\right)c_i)$,
15	$h\left(k\right)$ = GreedySearch2($h_i\left(k\right)$, $\mathcal{C}$),
16	End For
17	Output: $\mathit{h}=\left\{h\left(k\right)\right|h=0:N_c-1\}$
* Symbols in Algorithm 1 are described in the context of Section 3.1.