Recognizing Scrabble Letters

After labeling images of Scrabble letters on LabelBox, a Keras model is fed these images and labels. Then, predictions of future letters can be made.

Getting Started

You will need:

• Python 3.6
• Keras and Tensorflow
• NumPy
• OpenCV-python
• pandas
• scikit-image
• sklearn
• matplotlib.pyplot

File List

• letters.csv Downloaded labeled images of Scrabble letters from LabelBox
• loadimages.py Converts data (images) from letters.csv to a numpy array
• loadlabels.py Converts data (labels) from letters.csv to a numpy array
• resize_images.py Takes the numpy array from loadimages.py and resizes each element to 100 by 100 and makes them grayscale.
• main.py creates a Keras model for recognizing scrabble images
• labels.npy a NumPy file of the labels
• resized_data.npy a NumPy file of the resized and recolored images

Author

• Alexander Faus