Image Classifier

This is a simple image classifier built using Tensorflow that is attached to a Flask server and a Vue user interface that can upload images and receive results from the trained model. This model achieved 97% training accuracy and 67% validation accuracy.

To use the classifier

1. Put the sorted training images into their labeled folders in the training_images folder
2. Edit the training categories in line 19 of train.py
3. Run train.py
4. To predict, start server.py and open index.html

Note: By default, the server runs on port 5000. If nothing happens when you a submit an image, CORS is blocked the POST request.

The classifier contains

• 2 Convolutional layers
• 1 Flattening layer
• 1 Dropout layer
• 2 Dense layers