ImageNet

ImageNet is one of the largest annotated datasets for computer vision. The data is arranged according to a hierarchical order. There are 1,000 classes with 1.4 million images overall. Though the images are for non-commercial use, ImageNet is still one of the most popular datasets when it comes to learning computer vision. Especially in deep learning, the dataset is used to create image classification models due to availability of large number of varied images. 

The following website provides links and resources to download image URLs or other attributes about images:

http://image-net.org/download

In this book, ImageNet is not used explicitly, but we will be using a pre-trained model on it. There is no requirement to download this dataset for this book.