Sufficient to get full score for segmentations.
TODO:
- Få tak i RGB for pixel
- RGB-Distanse mellom piksler
- Putte pixel i segment-set(Lage genotype)
- Fitness for segment
- Find centroid(average)
- Find border-pixels
- For pixel find neighbour pixels
- dynamic? vector!
- make 4 & 8(n-boxses) neighbourhood
- Iterator for point_vec_t
- Find neighbour-segments of segment
- Check if neighbour
- Get all neighbours
- Initialize Solution
- Mutation
- Merge
- put segment together
- Check for conectivity
- Split
- Merge
- Crossover
- sort segment from best to worse
- [?] Fix "broken" segments
- Fitness for Solution
- overall deviation
- edge value
- connectivity measure