…These images are manually labelled and segmented according to a hierarchical taxonomy to train and evaluate object detection algorithms. The annotations are provided in COCO format.
10 PAPERS • NO BENCHMARKS YET