Table 2.

The information of the gait information introduction approach.

Name	Expression	O	Dynamic	Static	Time	Image
MIEI	Ei = α × Bi−1 + (1 − α) × Ei−1, α ∈ (0,1)	O(η·m·n)	√	√	√
FDEI	$\begin{array}{l}{E}_{\mathit{\text{FDEI}}}(x,y)=F(x,y,n)+E_{\mathit{\text{DEI}}}^c(x,y)\hfill \\ {E^c}_{\mathit{\text{DEI}}}(x,y)=\{\begin{array}{ll}\frac{1}{N_c}\text{∑}_{n\in N_C}B(x,y,n)\hfill & \mathit{\text{if}}\frac{1}{N_c}\text{∑}_{n\in N_C}B(x,y,n)\ge T\hfill \\ 0\hfill & \mathit{\text{otherwise}}\hfill \end{array}\hfill \\ F(x,y,n)=\{\begin{array}{ll}0\hfill & \mathit{\text{if}}B(x,y,n)\ge B(x,y,n-1)\hfill \\ B(x,y,n-1)-B(x,y,n)\hfill & \mathit{\text{otherwise}}\hfill \end{array}\hfill \end{array}$	O(η·m·n)	√	√	×
EGEI	$\begin{array}{l}{E}_{\mathit{\text{EGEI}}}(x,y)=G(x,y)\times {(T_{\mathit{\text{DWN}}}(x,y))}^{\gamma }\hfill \\ {\sigma }_{\mathit{\text{GEI}}}(x,y)=\sqrt{\frac{1}{A}\text{∑}_{i=1}^A{\left[\frac{1}{N}\text{∑}_{m=1}^N{G}_m^i(x,y)-\frac{1}{A}\text{∑}_{i=1}^A\frac{1}{N}\text{∑}_{m=1}^N{G}_m^i(x,y)\right]}^2}\hfill \end{array}$	O(A2·η·m·n)	√	√	×
CGI	$E_{\mathit{\text{CGI}}}(x,y)=\frac{1}{p}\text{∑}_{i=1}^p{\sum }_{t=1}^{n_i}{C}_t(x,y)$	O(p·ηi·(m·n)2)	√	√	√
GFI	$\begin{array}{l}{E}_{\mathit{\text{GFI}}}(x,y)=\frac{{\sum }_{n=1}^{N-1}{F}_n(x,y)}{N}\hfill \\ F_i(x,y)=\{\begin{array}{ll}0\hfill & \mathit{\text{if}}\mathit{\text{Mag}}{F}_i(x,y)\ge 1\hfill \\ 1\hfill & \mathit{\text{otherwise}}\hfill \end{array}\hfill \\ \mathit{\text{Mag}}{F}_i(x,y)=\sqrt{{(\mu F_i(x,y))}^2+{(\nu F_i(x,y))}^2}\hfill \end{array}$	O(η·(m·n)2)	√	×	√
GEnI	$E_{\mathit{\text{GEnI}}}(x,y)=-E_{\mathit{\text{GEI}}}(x,y){log}_2{E}_{\mathit{\text{GEI}}}(x,y)-(1-E_{\mathit{\text{GEI}}}(x,y)){log}_2(1-E_{\mathit{\text{GEI}}}(x,y))$	O(η·m·n)	√	√	×

Note: We make √ and × represent whether the Class Energy Image with or without the type of the motion information, respectively. O denotes computational complexity. Suppose the size of the silhouette is m×n, η is the numbers of silhouettes in a gait cycle.