Long Video Dataset

We randomly selected three videos from the Internet, that are longer than 1.5K frames and have their main objects continuously appearing. Each video has 20 uniformly sampled frames manually annotated for evaluation.

Benchmarks

Semi-Supervised Video Object Segmentation/J&F Semi-Supervised Video Object Segmentation/J Semi-Supervised Video Object Segmentation/F Video/J&F Video/J Video/F Video Object Segmentation/J&F Video Object Segmentation/J Video Object Segmentation/F

Related Benchmarks

Long Video Dataset (3X)/Semi-Supervised Video Object Segmentation/F Long Video Dataset (3X)/Semi-Supervised Video Object Segmentation/J Long Video Dataset (3X)/Semi-Supervised Video Object Segmentation/J&F Long Video Dataset (3X)/Video/F Long Video Dataset (3X)/Video/J Long Video Dataset (3X)/Video/J&F Long Video Dataset (3X)/Video Object Segmentation/F Long Video Dataset (3X)/Video Object Segmentation/J Long Video Dataset (3X)/Video Object Segmentation/J&F