predrag-njegovanovic / poker-hand-recognition

Classification and recognition of poker hands via OpenCV and machine learning algorithms

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

PokerHandRecognition

Description of project

Recognition of playing cards and classification of poker hands.

Libraries necessary to run scripts:

-python 2.7
-opencv 3.2
-numpy 1.12 or greater
-sckit-learn 0.18
-sckit-image 0.10
-h5py
-keras
-Theano or Tensorflow

All libraries can be installed via pip. Run command pip install <library_name>.

Run main.py -> Load models and recognize all cards from folder 'data/test_dataset/' and write result in /results/results.txt. If models aren't saved run train_nn.py first.

card_manipulation.py -> Script contains functions for image processing

crop_cards.py and preprocess_crop_cards.py -> Create dataset to run through neural network

train_nn.py -> Create models for NN and Random Forest and save to file

About

Classification and recognition of poker hands via OpenCV and machine learning algorithms

License:MIT License


Languages

Language:Python 100.0%