Given an image, Visual Relationship Detection involves detecting objects shown and inferring relationships (semantic, spatial etc.) between them. Click on one of the example images or run on your own images! For more details you can visit our API documentation.