Pattern Recognition?

Discussion in 'Game Development (Technical)' started by Davaris, Aug 20, 2007.

  1. bvanevery

    bvanevery New Member

    Joined:
    Jun 25, 2007
    Messages:
    369
    Likes Received:
    0
    I'm not convinced that a correct game AI problem has even been stated. This smacks too much of "I want the AI to figure everything out for me."
     
  2. voxel

    voxel New Member

    Joined:
    Oct 29, 2006
    Messages:
    176
    Likes Received:
    0
    You have to reduce the problem:

    Image -> smaller data (aka feature vectors)

    The idea is to convert the image into a different smaller representation - i.e generating an outline of the Y by running an edge detection algorithm (convolution) over it and say generating a series of connected lines.

    The recognition is done on the feature vectors not the original image.

    In speech you have:

    Sound -> phonemes

    which is similar to the above.
     
  3. RWgames

    Original Member

    Joined:
    Jan 26, 2006
    Messages:
    24
    Likes Received:
    0
    Hi, I haven't read through every post, so sorry if it's been mentioned before, but I recommend you consider the book 'AI Techniques for Game Programming' by Mat Buckland and/or see his website. Nice, and easy to digest, information on neural networks and such. :)
     
  4. FlySim

    Indie Author

    Joined:
    Jul 30, 2004
    Messages:
    352
    Likes Received:
    0
    I suspect you could solve this problem using image pattern and/or feature matching techniques. But Im not sure it worth the trouble unless this will REALLY make a difference in game play. I use some of these techniques for some of my contract gigs and its fairly complicated software.

    Here's what I consider to be the state of the art feature matching software:

    http://www.cs.ubc.ca/~lowe/keypoints/
     
  5. Jesse Aldridge

    Original Member

    Joined:
    Apr 19, 2005
    Messages:
    221
    Likes Received:
    0
    I started to do gesture recognition for one of my games. It actually worked pretty well. It basically goes like this:

    Manually create a list of points that represents what the pattern should look like - I'll call this the test list.

    Store the mouse position at some interval (say, once per cycle) in a list of points - I'll call this the point list.

    Loop through each point in the test list and calculate the distance between it and it's corresponding point in the point list. Add the distances together to find the "offness".

    The only somewhat tricky part is matching the points in the test list to the corresponding point in the point list. Just visualize cutting the point list into n equal portions where n = the number of points in the test list and then taking the point at each "cut". There's probably a better way to explain that mathematically, but math really isn't my thing :)
     

Share This Page

  • About Indie Gamer

    When the original Dexterity Forums closed in 2004, Indie Gamer was born and a diverse community has grown out of a passion for creating great games. Here you will find over 10 years of in-depth discussion on game design, the business of game development, and marketing/sales. Indie Gamer also provides a friendly place to meet up with other Developers, Artists, Composers and Writers.
  • Buy us a beer!

    Indie Gamer is delicately held together by a single poor bastard who thankfully gets help from various community volunteers. If you frequent this site or have found value in something you've learned here, help keep the site running by donating a few dollars (for beer of course)!

    Sure, I'll Buy You a Beer