Sign in to follow this  

Simplest pattern recognition

This topic is 3855 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

I'm not sure if this should be in general programming. I have N (3~6) spheres in 3D space connected by lines. Each line has a strength assigned (1-4). The shorter the distance, the bigger the strength. I need a process that recognizes shapes. For example: - A triangle where all sides are connections of the same strength. - A rhombus where all sides are connections of the same strength as the shortest of the central connections. - A regular pentagon (the weak connections inside would form a pentagram). Could (and should) I solve this with an AI or I should move the thread to general programming. Thanks.

Share this post


Link to post
Share on other sites
Since your classes are so well defined, I would use a simple classification tree, using such parameters as the number of connections, the connection strengh span, etc.

Share this post


Link to post
Share on other sites
Aye, I'd agree that a decision/classification tree would be a good way to go with this. You've got a very constrained number of vertices (spheres) and a dependent number of edges (lines) and a small enumeration of the possible shapes (leaf nodes in the tree).

Share this post


Link to post
Share on other sites
You were right.

The items are left free in a 3D space and they build the shapes following an attraction/repulsion function.

Five items can form:
A- a 'C'.
B- a square with an item in the center.
C- a 'W'.
D- a pentagon.
E- a bird foot.

Listing the "hard connections" (the ones that form between closest items)

1) 1, 1, 2, 2, 2
2) 2, 2, 2, 2, 4
3) 2, 2, 3, 3, 4
4) 2, 2, 2, 2, 2
5) 1, 2, 2, 3, 4

I'll just have to do this for groups of 4, 5 and 6.


Thanks for pointing in the right direction, I was about to follow a much longer path. :)

Share this post


Link to post
Share on other sites

This topic is 3855 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this