Selection type:
elitist
tournament
Elite length (how many individuals with pass untouchable to the next generation):
Tournament size:
Operators:
Mutation and Crossover
Mutation
Crossover
— Crossover:
Semantic crossover (a point crosses over with the closest point of another individual rather then shift
individual's
slices)
Probability for doing semantic crossover rather then normal crossover when semantic crossover is on (its
off)
Probability for performing crossover
Probability for crossing over with closer points (the closer the more probable):
Probability for crossing over with initial individuals rather than the other tournament winner:
Maximum percentage of points to crossover in one iteration:
— Mutation:
Semantic mutation (select a slice of points and move them all accordingly to a vector rather then randomly
vary each point)
Probability for performing mutation:
Probability for each point to be moved when mutating:
Maximum distance for a point to move when mutating:
Probability for deleting a point when mutating:
Probability for adding a point when mutating:
Probability for the added point to be of the type "L" rather then "M":
Probability for changing the type of one point when mutating:
Maximum percentage of points to mutate in one iteration:
Evolutionary process:
Maximum iterations:
Target confidence for a glyph to represent the character:
Satisfactory distance to the target confidence:
Minimum size (percentage of the canvas) that a glyph must have to be considered a valid character:
Stops if it finds this percentage of fitted individuals:
Other options:
Draw red rects over the detection zone
Download files during auto evolution
The the aforementioned setup this times (useful if downloading)