Algorithm 1 Algorithm of SS |
Input: An input image I with height H and width W
Output: The final saliency map S
1: sumSal = Zeros(H,W); k = 0;
2: for scale = 1 to s do
3: for each CIELAB color channel do
4: k = k + 1
5: Set width or height of image border as δ = scale;
6: Associate all pixels in the four borders with a strength of one;
7: Iterate using Equations (3) and (4) until each pixel Pi keeps an unchanged strength S(Pi);
8: Get a full-field saliency map salMap[k], each pixel Pi of which is assigned with the saliency value 1 − S(Pi);
9: sumSal = sumSal + salMap[k];
10: end for
11: end for
12: S = L2Normalization(sumSal);
|