To perform universal event stream segmentation, we collected a large-scale RGB-Event dataset for event-centric segmentation, from current available pixel-level aligned datasets (VisEvent, COESOT), namely RGBE-SEG. The RGBE-SEG included 65,957 image-event pairs, 64,957 for training and 1,000 for testing. The test set contained 38,760 masks, and we artificially divided it into easy, medium, and hard subsets based on the complexity of scenarios. All ground truth masks were generated by images and the well-trained SAM.
Variants: RGBE-SEG
This dataset is used in 1 benchmark:
| Task | Model | Paper | Date |
|---|---|---|---|
| Event-based Object Segmentation | EventSAM | Segment Any Events via Weighted … | 2023-12-24 |
| Event-based Object Segmentation | SAM | Segment Anything | 2023-04-05 |
| Event-based Object Segmentation | ESS | ESS: Learning Event-based Semantic Segmentation … | 2022-03-18 |
| Event-based Object Segmentation | Evdistill | EvDistill: Asynchronous Events to End-task … | 2021-11-24 |
| Event-based Object Segmentation | DTL | Dual Transfer Learning for Event-based … | 2021-09-04 |
| Event-based Object Segmentation | NGA | Learning to Exploit Multiple Vision … | 2020-03-24 |
| Event-based Object Segmentation | E2VID | High Speed and High Dynamic … | 2019-06-15 |
Recent papers with results on this dataset: