A convolutional neural net that classifies images into two categories: corgis or bread.
Inside this zip file, you will find:
- (Constants.m) A file with constants that are used throughout the project.
- (preprocessImages.m) A file that resizes images and catches corrupted/non-RBG images.
- (setupCNN.m) A file that setups the convolutional neural network.
- (displayMislabeledImages.m) A file that creates a figure of the first 100 mislabeled images.
- (CorgiBreadClassifier.m) And finally, a script that runs everything up above!
To build this project, please fill in your directory constant (in Constants.m) where you downloaded this zip file/where the data can be found on your computer. Then, replace the root_folder variable on line 2 (in CorgiBreadClassifier.m) with your newly created constant. From here, all you have to do is run CorgiBreadClassifier.m!
The script will train a CNN, outputting its accuracy and displaying the first 100 mislabeled images.
Enjoy, Emily and Steph 🎉