Algorithm 1 Past events remove mechanism
Input: Spatiotemporal event stream: ${\displaystyle \sum }_{i=1}^{N}e\left(x_i,y_i,t_i\right)$
Output: An event stream that can form a reference frame with clear and sharp edges: $E\left(\Delta f\right)={\displaystyle \sum }_{t_f}^{t_k+\Delta f}\{e\left(x_i,y_i,t_i\right)|t_i\in \left[t_f,t_k+\Delta f\right]\}$
1	For $e\left(x_i,y_i,t_i\right)$ in ${\displaystyle \sum }_{i=1}^{N}e\left(x_i,y_i,t_i\right)$ do
2	$\hspace{1em}R=\{e\left(x,y,t\right)|x\in \left[x_i-1,x_i+1\right],y\in \left[y_i-1,y_i+1\right],t\le t_i\}$
3	Calculating the optical flow information of R
4	Get $\overrightarrow{v_{\left(e_i\right)}}$ of $e\left(x_i,y_i,t_i\right)$ by vector synthesis
5	For $e\left(x,y,t\right)$∈R do
6	2, 3, and 4
7	End
8	Get $\overrightarrow{V_{e_i}}$ of $e\left(x_i,y_i,t_i\right)$ by local consistency
9	Obtain the past events of $e\left(x_i,y_i,t_i\right)$ according to the movement direction of the event, and remove the past events.
10	End