Download presentation
Presentation is loading. Please wait.
Published byDenis Alan Barker Modified over 8 years ago
1
Bara Lilla Nyíri Gergely Piotr Czekański Kovács Laura Team H: Automatic Poker Player
2
Automatic Poker Player
3
Usage Determine the shapes of poker-cards (i.e. the hand value) Difficulties: Hidden parts of cards Cards in different positions (angles) Motivation: Electronic Casinos
4
General Presentation Corners of the cards Pattern recognition Image Processing - method 1 - Existing symbols Image Processing - method 2 -
5
Detailed Presentation Image Processing – method 1 Image Processing – method 2 Pattern recognition
6
Image Processing – method 1 Step 1: Tresholding using isodata algorithm Step 1: Tresholding using isodata algorithm Step 2: Fill area, closing holes Step 2: Fill area, closing holes Step 3: Determine the boundary Step 3: Determine the boundary Step 4: Labeling the boundaries Step 4: Labeling the boundaries Step 5: Compute the chain code Step 5: Compute the chain code Step 6: Determine the corners Step 6: Determine the corners Step 7: Determine the angle Step 7: Determine the angle Step 8: Rotate the cards Step 8: Rotate the cards Step 9: Pattern matching Step 9: Pattern matching
7
Isodata algorithm divide the histogram in two parts (starting treshold level: t = the mean of the picture) consider the means (m1, m2) of the upper and lower pixelvalues Calculate: t= (m1+m2) /2 If t changes => start from the beginning Step 2, otherwise (i.e. convergence)
8
Fill area, closing holes Consider a closed contour and an inner point => Dilatation => intersection with the complementary of the contour => Continue until there are changes
9
Image Processing – method 2 Localize the cards Keep only the symbols Extracting the relevant symbols – from distance analysis Pattern recognition – numbers and symbols
10
Pattern Matching
11
Conclusions Results Further Work
12
Results Detecting the shape of poker cards - even if some parts are hidden The performances of the algorithms depend on the numbers of the contained symbols
13
Problems
14
Further Work Using last year’s Dice Project implementing an SSIP Casino Analyze the information of the shapes => determine the hand value of the poker cards Improve the algorithms by using fuzzy logic or neural networks
15
Thank You for the Attention!
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.