Uncommon Objects

Uncommon Objects challenges the limitations of machine vision by subverting traditional expectations of object recognition. The project employs the COCO (Common Objects in Context) dataset to detect objects in images, but inverts the usual process: objects identified with high confidence are blurred, while those detected with low confidence—or not detected at all—remain clear. This approach draws attention to what machines struggle to perceive, exposing the blind spots and failures of artificial intelligence systems. By revealing how models are shaped by their training data–by these definitions of 'common objects'–the project exposes the gaps that emerge from those definitions and datasets.