Image Classification using Convolutional Neural Networks

This project aims to classify the images in the given dataset as cats or dogs using convolutional neural networks(CNN)

Approach and pipeline:

Refer to the report and code for the approach and implementation.

Results:

• Results after training 18,000 images of cats and dogs:

  • number of epochs = 15
  • training data / validation data split = 80/20
  • MODEL
    • CONV 3x3 filter layers with batch norm - 32 x 64 x 96 x 96 x 64
    • Dense layers with drop out of 0.2 and 0.3 - 256 x 128 x 2
    • loss: 0.0638
    • accuracy: 0.9759
    • val_loss: 0.3255
    • val_accuracy: 0.9044

• The model was tested on the images in the test1 folder. The performance of the model was very good and was able to predict the animals with 97-99% accuracy.

Plots for model accuracy and loss are following:

Classifying the images:

Output video predicting the images as cats and dogs can be found here.

Instructions to run the code:

Input dataset

• Go to directory: cd Code/
• To start the training run:
  • $ python main.py