AMT Objects

AMT Objects is a large dataset of object centric videos suitable for training and benchmarking models for generating 3D models of objects from a small number of photos of the objects. The dataset consists of multiple views of a large collection of object instances.

The dataset contains 7 object categories from the MS COCO classes: apple, sandwich, orange, donut, banana, carrot and hydrant. For each class, annotators were asked to collect a video by looking ‘around’ a class instance, resulting in a turntable video. The dataset contains 169-457 videos per class. For each class, the videos were randomly split into training and testing videos in an 8:1 ratio.