Package edu.uky.ai.ml

Contains an example exercise in which a neural network is trained to recognize images of capital letters.

This example provides an interactive GUI which allows you to draw simple 5x7 pixel images in black and white. It also provides a database of images corresponding to the capital letters in the Roman alphabet. A simple feed-forward neural network can be trained, using the back propagation algorithm, to correctly recognize the character given the image.

You should start by familiarizing yourself with the code used in the neural networks package. Then fill in the following methods in this order:

  • Class Summary 
    Class Description
    AlphabetDatabase
    A training database of 5x7 pixel images of capital letters.
    AlphabetFrame
    An interactive GUI that allows you to interact with a neural network that has been trained to recognize images of capital letters.
    AlphabetInputPanel
    A 5x7 array of toggelable pixels that allows the user to draw an image which the neural network will attempts to classify.
    Main
    Creates and trains a neural network to recognize capital letters.