Open Call: Sam Lavigne

Sam Lavigne’s Training Poses explores how machines see bodies. Machine learning systems require enormous datasets to learn how to see. One such dataset is Microsoft’s COCO (Common Objects in Context), an archive of 220,000 images originally found on Flickr. The images in COCO have been tagged and annotated by workers on Amazon Mechanical Turk, an online crowdsourcing marketplace. Countless machine learning systems use COCO to enable object recognition technology. In Training Poses, participants are invited to encounter this otherwise-invisible dataset by matching their poses with a random selection of images from COCO.