Algorithm 1 Pseudo-code of location refinement matching
1:	Inputs: Test image I, size of template w × h, coordinate candidate ($x_{Cand1}$, $y_{Cand1}$), angle
	candidates ${\theta }_{Cand1},$ scale candidates $s_{Cand1}$, pyramid level $N_P$, number of candidates $N_{Cand1}$
2:	Outputs: Correlation coefficient ${\eta }_{NCC}$
3:	X index: idxX$\leftarrow$blockDim.x ∗ blockIdx.x + threadIdx.x   //Number of extension pixels
4:	Y index: idxY$\leftarrow$blockDim.y ∗ blockIdx.y + threadIdx.y   //Number of candidates
5:	if (idxX < $2^{N_P}$) and (idxY < $N_{Cand1}$) then
6:	Coordinate X: $x_{Refined}\leftarrow$(idxX mod $2^{N_P}$) + ($x_{Cand1}$ + $2^{N_P}$)  //Refinement coordinate x
7:	Coordinate Y: $y_{Refined}\leftarrow$ (idxY div $2^{N_P}$) + ($y_{Cand1}$ + $2^{N_P}$)  //Refinement coordinate y
8:	Scale s: s $\leftarrow s_{Cand1}$[idxY]
9:	Angle $\theta$: $\theta$ $\leftarrow {\theta }_{Cand1}$[idxY]
10:	for j in h do
11:	for i in w do
12:	Collect intensity pixel values inside a search window with a center point at
	($x_{Refined}$, $y_{Refined}$), an orientation: $\theta$, and a scale: s
13:	end
14:	end
15:	Calculate the NCC score ${\eta }_{NCC}$ between the template and the search windowend
16:	end
17:	Return${\eta }_{NCC}$