We are in the streets everyday, all over the world, constantly taking new photographs of products and objects.
Our photos show what's out there right now, what's hot and what's not.
• Significant differences between images.
• Datasets optimised for machine learning
We manually create tight bounding boxes.
Bounding boxes are annotated with additional information such as whether the tag is the subject or the background.
Pixel perfect object extraction, painstakingly created by hand.