Skip to content

ColeBardin/Cheeser

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

106 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Cheeser

Is this cheese?

Image recognition software that trains to detect images of cheese using SciKit-Learn models

Trains with images of cheese and not cheese then makes predictions on a fraction of the sample photos

cheeser.py

Usage: py cheeser.py [init|load|test]

Train software with photos of cheese and photos of not cheese

Use the directores data\Cheese\ and data\NotCheese\ to train the AI

Optional init commandline argument will reload all the images into the .pkl file before training. This only needs to be run if there new photos have been added. Without the init flag, cheeser.py will attempt to read from a .pkl file composed of the base name and the dimensions

Optional load flag will not train a new HOG SGD model, but instead read one in from hog_sgd_model.pkl

Optional test flag will read in testing data from indir\. When called, the program will try to find a fully trained model under the name full_train_model.pkl. If does not exist, it will train one with all the images in data\ and create a .pkl file for it

Specifying no flags will split the data from data\ to 90% for training and 10% for testing. Then it will create a CLF and Grid Search model, compare the two and display the results. Lastly, the program will give you the option to save the trained model as hog_sgd_model.pkl

Dependencies

cheeser.py:

Matplotlib

Joblib

Numpy

Scikit-Learn

Scikit-Image

transformer_classes.py:

Numpy

Scikit-Learn

Scikit-Image

About

Is this cheese?

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages