Research Article
Initial Geometrical Templates with Parameter Sets for Active Contour on Skin Cancer Boundary Segmentation
Algorithm 1
Template selection in training state.
| (1) | START | | (2) | SET Tp (i): candidate templates (e.g., circle, rectangle, ellipse, triangle, etc.), n: number of candidate templates, : template rotation angle, : offset of rough skin cancer image | | (3) | REPEAT | | (4) | INPUT an image of skin cancer | | (5) | COMPUTE binarization | | (6) | COMPUTE rough contour of a skin cancer image | | (7) | COMPUTE scaling the skin cancer image by offset () | | (8) | COMPUTE centroid and the longest radius of skin cancer | | (9) | REPEAT | | (10) | INPUT candidate template () | | (11) | COMPUTE centroid and the longest radius of template | | (12) | COMPUTE mapping between candidate template and skin cancer image based on centroid and the longest radius | | (13) | COMPUTE scaling the template based on the longest radius of skin cancer image with offset | | (14) | SET | | (15) | REPEAT | | (16) | COMPUTE difference between candidate template () and skin cancer image | | (17) | CALCULATE | | (18) | COMPUTE rotation of candidate template on the centroid | | (19) | UNTIL all rotation angles () | | (20) | SELECT angle with the smallest difference | | (21) | CALCULATE | | (22) | UNTIL all candidate templates | | (23) | SELECT template and angle with the smallest difference with the skin cancer image | | (24) | UNTIL all skin-cancer image samples | | (25) | END |
|
