Information / AI & Robotics

ImageNet dataset

Fei-Fei Li's team built ImageNet (2009): millions of labelled web photos that gave deep networks the data they needed.

From 2007 Fei-Fei Li and colleagues collected millions of images from the web and had them labelled by crowd workers on Amazon Mechanical Turk, presenting ImageNet at the CVPR conference in 2009. Its yearly challenge, from 2010, compared methods on about 1.2 million training images in 1,000 classes. When AlexNet won it by a wide margin in 2012, it showed that large data plus GPUs beat hand-built features.