Image Classification with A CNN (Tensorflow)

This is a basic CNN model built through Python Keras/Tensorflow to identify images of cats and dogs. You can view the code and try it for yourself here!

The dataset above was taken from kaggle and provided thousands of photos of both cats and dogs to use for the classification.

#2. Image prep

Images were put into greyscale and at a fairly low resolution as both color and resolution were both considered of not great importance when comparing cats and dogs. This reduces runtime and memory storage.

Convolutional neural networks are typically used for image classification. Using tensorflow and a typical CNN layout to begin with, I took advantage of tensorboard and ran multiple models under differing NNs (as can be seen below). My overal best run model held an accuracy of 82%.