Algorithm 1.

GetBBM_twosamples.

INPUT: LBPCode1 of Img1, LBPCode2 of Img2, (moving) step, (moving) times.

OUTPUT: BBM of two samples Img1 and Img2.

Algorithm:

Direct matching, compute the similarity using LBPCode1and LBPCode2, denoted as S0.

//taking Img1as base, align through moving Img2 toward up direction.

for n=1 to times do

move Img2 up n*step pixels.

extract the subsets of LBPCode1 and LBPCode2 according overlapped region.

Compute the similarity using these two subsets, denoted as Sup(n).

endfor

//taking Img1as base, align through moving Img2 toward down direction.

for n=1 to times do

move Img2 down n*step pixels.

extract the subsets of LBPCode1 and LBPCode2 according overlapped region.

Compute the similarity using these two subsets, denoted as Sdown(n).

endfor

//taking Img1as base, align through moving Img2 toward left direction.

for n=1 to times do

move Img2 left n*step pixels.

extract the subsets of LBPCode1 and LBPCode2 according overlapped region.

Compute the similarity using these two subsets, denoted as Sleft(n).

endfor

//taking Img1as base, align through moving Img2 toward right direction.

for n=1 to times do

move Img2 right n*step pixels.

extract the subsets of LBPCode1 and LBPCode2 according overlapped region.

Compute the similarity using these two subsets, denoted as Sright(n).

endfor

S=Max(S0,Sup(1)…Sup(times),     Sdown(1)…Sdown(times),     Sleft(1)…Sleft(times), Sright(1)…Sright(times))

Get the subsets of LBPCode1and LBPCode2 corresponding to S.

Take the subsets as the new LBPCodes.

generate the BBM of new LBPCodes using definition 1 and 2.