From: Co-occurrence context of the data-driven quantized local ternary patterns for visual recognition
Mapping table generation algorithm | |
---|---|
Input: Number of neighbor pixel L | |
Output: Mapping table M (N P ×N P matrix) | |
Initialization: Index=1, N P =3L, \(\mathbf {M}\Leftarrow \{null\}^{N_{P} \times N_{P}}\) | |
for: p=0,⋯,N P −1 and q=0,⋯,N P −1 | |
if M(p,q)=n u l l, then | |
M(p,q)⇐I n d e x, p ′⇐s h i f t((p)3,2), | |
q ′⇐s h i f t((q)3,2), M(q ′,p ′)⇐I n d e x | |
p ′⇐s h i f t((p)3,1), | |
q ′⇐s h i f t((q)3,1), M(p ′,q ′)⇐I n d e x | |
p ′⇐s h i f t((p)3,3), | |
q ′⇐s h i f t((q)3,3), M(q ′,p ′)⇐I n d e x | |
I n d e x⇐I n d e x+1 | |
end if | |
end for |