CNNs to improve accuracy in the case of image translation

In the previous sections, we learned about the issue of translation in images and how a CNN works. In this section, we will leverage that knowledge to learn how a CNN works toward improving prediction accuracy when an image is translated.

Getting ready

The strategy that we will be adopting to build a CNN model is as follows:

• Given that the input shape is 28 x 28 x 1, the filters shall be 3 x 3 x 1 in size:
  • Note that the size of filter can change, however the number of channels cannot change
• Let's initialize 10 filters
• We will perform pooling on top of the output obtained in the previous step of convolving 10 filters over the input image:
  • This would result...