The dataset used in the paper is the Pascal VOC 2012 dataset, which is a benchmark for instance segmentation. The dataset consists of 1464 images with 20 class categories and 2.8 objects per image.
BibTex:
Before browse our site, please accept our cookies policy
