Skip to content

A university project to apply a multilayer perceptron network onto an algorithm, so that it can learn to identify the classes of some images (used on the training process). When you provide an image for it to judge, you will be told the probability of this image to be each of the trained classes.

Notifications You must be signed in to change notification settings

gucollaco/drawing-identifier

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

drawing-identifier

A university project to apply a multilayer perceptron network onto an algorithm, so that it can learn to identify the classes of some images (used on the training process). When you provide an image for it to judge, you will be told the probability of this image to be each of the trained classes.

Firstly, check if you have the following libraries installed (and if possible, update them):

  • Keras
  • NumPy
  • Scikit-learn
  • Matplotlib
  • PIL

Then, select on Quick, Draw!, some .npy data to be used on the experiment. Once you selected the desired datas, create a 'data' folder on the root of the project, and put them inside it (the code will use these files on the learning process).

The next step is to execute the main.py file, which will accomplish the training process, and store a keras model at the 'model' folder.

python main.py

Once this is done, run the verify.py file, which will check if the image you want to classify (you can use one of the images on the 'my_drawings' folder), and plot a chart containing the information seeked.

python verify.py

About

A university project to apply a multilayer perceptron network onto an algorithm, so that it can learn to identify the classes of some images (used on the training process). When you provide an image for it to judge, you will be told the probability of this image to be each of the trained classes.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages