fashion_mnist

1. Dataset Description

• Source: https://www.kaggle.com/zalando-research/fashionmnist
• Size: 60,000 training set and 10,000 test set of 28x28 grayscale images
• Labels:
  • 0 - T-shirt/top
  • 1 - Trouser
  • 2 - Pullover
  • 3 - Dress
  • 4 - Coat
  • 5 - Sandal
  • 6 - Shirt
  • 7 - Sneaker
  • 8 - Bag
  • 10 - Ankle boot

2. Preprocessing

• Image equalization
• Grayscale to binary

3. Predictive Model

• Building CNN layers

4. Visualize filters for each layer

Filter layers

5. Visualize activation layer for test set

Test image (binary):

Activation layers: