I like the idea of classifying trees by looking at the parameters of the branch structure -- I think that kind of approach would very well complement traditional computer vision methods.
Another thing that I've been wanting to play with is creating a 'forest' of these kinds of parametrizable trees and running a genetic algorithm to see what parameters win out in collecting "sunlight", etc. Could be fun, but would take a bit longer to make...
I like the idea of classifying trees by looking at the parameters of the branch structure -- I think that kind of approach would very well complement traditional computer vision methods.
Another thing that I've been wanting to play with is creating a 'forest' of these kinds of parametrizable trees and running a genetic algorithm to see what parameters win out in collecting "sunlight", etc. Could be fun, but would take a bit longer to make...