To facilitate research on understanding objects in 3D, Google Research has released Objectron – a dataset of 15,000 annotated video clips and 4M annotated images of objects captured from different angles. Each sample also includes AR session metadata on object and camera pose, sparse point clouds, and a characterization of the planar surfaces in the surrounding environment. The dataset, which includes objects like bikes, books, bottles, cameras, and chairs, was collected from 10 countries on five continents to ensure geo-diversity.