ImageNet

ImageNet is a large dataset containing more than 14 million images annotated for image classification or object detection. It was first consolidated by Fei-Fei Li and her team in 2007. The goal was to build a dataset that computer vision researchers could benefit from.

The dataset was presented for the first time in 2009, and every year since 2010, an annual competition called the ImageNet Large-Scale Visual Recognition Challenge (ILSVRC) has been organized for image classification and object detection tasks.

Figure 8.1: Examples of images from ImageNet

Over the years, some of the most famous CNN architectures (such as AlexNet, Inception, VGG, and ResNet) have achieved amazing results in this ILSVRC competition. In the following graph, you can see how some of the most famous CNN architectures performed in this competition. In less than 10 years, performance increased from 50% accuracy to almost 90%.

Figure 8.2: Model benchmarking...