Hands-On Transfer Learning with Python

Dog Breed Identification dataset

In the previous section, we utilized a low-resolution image dataset to classify images amongst 10 non-overlapping categories. It was by no means a simple task, yet we achieved decent performance with minimal effort.

Now let us level up the game and make this task of image classification even more exciting. In this section, we will be concentrating toward the task of fine-grained image classification. Unlike usual image classification tasks, fine-grained image classification refers to the task of recognizing different subclasses within a higher-level class.

To help understand this task better, we will be focusing our discussion around the Stanford Dogs dataset (http://vision.stanford.edu/aditya86/ImageNetDogs/ ...