SBCoseg Dataset
The SBCoseg dataset includes 889 groups of images and each group consists of 18 images with a common object, leading to 16002 images in total. The whole dataset is divided into five subsets: with ECFB, with TR, with MH, with SD, and Normal (normal data). The five subsets contain 193, 251, 82, 83, and 280 image groups, respectively. Each original image is in JPG format with a pixel size of 360 ×360, and each ground-truth image is in PNG format.
Variants: SBCoseg
This dataset is used in 1 benchmark:
| Task | Model | Paper | Date |
|---|---|---|---|
| Semantic Segmentation | Dice loss + IS-Triplet loss | Improving Image co-segmentation via Deep … | 2021-03-19 |
Recent papers with results on this dataset: