Input: random path through all ROIs (path0), labels indicating which ROI corresponds to each path0 pixel (tags) |
Step 1. Find unique ROIs |
For ROI in unique(tags): Â Â Â Find ROI centroid |
Step 2. Find optimal path through ROIs centroids (pathTSP) applying genetic algorithm to solve TSP |
Randomly initialize a population of 100 individuals (random paths through all ROIs) Find minimum path length minglob and the individual than minimize path length (pathTSP) For generation in [1:1000]:    Compute the path length for each individual    If exist a path shorter then minglob, then update minglob and pathTSP    Randomly split population in groups of 4 individuals    For each group of 4 individuals:      Find the individual with shortest path      Create 4 mutations (original, flip, swap, slide) of the individual for the next generation |
Step 3. Compute pathSLS |
Initialize empty trajectory pathSLS |
For ROI in pathTSP:    Find path through all ROI’s pixels (pathROI) using a greedy algorithm    Append pathROI to pathTSP |
Return: pathTSP |