Saliency Mapping Enhanced by Structure Tensor

Algorithm 1

Algorithm of contour activation map.

(1) Input:

(2) Input image I: three-channel and size (, )

(3) Output:

(4) Edge feature map A: one channel and size (, )

(5) Corner feature map B: one channel and size (, )

(6) Contour activation map C: one channel and size (, )

(7) Begin

(8) Resize the input image I to (, ), called Im-Re

(9) for  ,   do

(10) for  ,   do

(11)  For Im-Re, calculate structure tensor using (6)

(12)  Calculate eigenvalues and using (7) and (8), respectively

(13)  

(14)  

(15) end for

(16) end for

(17) Normalize A and B into a fixed range 0⋯1

(18) Combine normalized A and normalized B into CT

(19) End