The dataset consists of 313 images of shirts. There are three photos of each product in a different position.
Labeling: Bounding boxes.
Metadata: Product category.