RefineVIS: Video Instance Segmentation with Temporal Attention Refinement
We introduce a novel framework called RefineVIS for Video Instance Segmentation (VIS) that achieves good object association between frames and accurate segmentation masks by iteratively refining the representations using sequence context. RefineVIS learns two separate representations on top of an off-the-shelf frame-level image instance segmentation model: an association representation responsible for associating objects across frames and a segmentation representation that produces accurate segmentation masks. Contrastive learning is utilized to learn temporally stable association representations. A Temporal Attention Refinement (TAR) module learns discriminative segmentation representations by exploiting temporal relationships and a novel temporal contrastive denoising technique. Our method supports both online and offline inference. It achieves state-of-the-art video instance segmentation accuracy on YouTube-VIS 2019 (64.4 AP), Youtube-VIS 2021 (61.4 AP), and OVIS (46.1 AP) datasets. The visualization shows that the TAR module can generate more accurate instance segmentation masks, particularly for challenging cases such as highly occluded objects.
PDF Abstract
Datasets
MS COCO
YouTube-VIS 2019
OVIS
Results from the Paper
Ranked #3 on
Video Instance Segmentation
on YouTube-VIS 2021
(using extra training data)
| Task | Dataset | Model | Metric Name | Metric Value | Global Rank | Uses Extra Training Data |
Benchmark |
|---|---|---|---|---|---|---|---|
| Video Instance Segmentation | OVIS validation | RefineVIS (Swin-L, offline) | mask AP | 46 | # 8 | ||
| AP50 | 70.4 | # 7 | |||||
| AP75 | 48.4 | # 7 | |||||
| AR1 | 19.1 | # 7 | |||||
| AR10 | 51.2 | # 5 | |||||
| Video Instance Segmentation | YouTube-VIS 2021 | RefineVIS (Swin-L, online) | mask AP | 61.4 | # 3 | ||
| AP50 | 84.1 | # 2 | |||||
| AP75 | 68.5 | # 3 | |||||
| AR10 | 65.2 | # 4 | |||||
| AR1 | 48.3 | # 5 | |||||
| Video Instance Segmentation | YouTube-VIS validation | RefineVIS (Swin-L, offline) | mask AP | 64.4 | # 8 | ||
| AP50 | 88.3 | # 3 | |||||
| AP75 | 72.2 | # 5 | |||||
| AR1 | 55.8 | # 8 | |||||
| AR10 | 68.4 | # 8 |
