Panoptic-PartFormer

Reported on 2 benchmarks across 1 task · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision2 results

Part-aware Panoptic SegmentationonCityscapes Panoptic Parts
PartPQ· 2022-04-10
61.9
best: 64.8 (TAPPS (Swin-B, COCO pre-training))
SOTA
Panoptic-PartFormer: Learning a Unified Model for Panoptic Part Segmentation arXiv:2204.04655
Part-aware Panoptic SegmentationonPascal Panoptic Parts
PartPQ· 2022-04-10
39.3
best: 60.4 (TAPPS (Swin-B, COCO pre-training))
SOTA
Panoptic-PartFormer: Learning a Unified Model for Panoptic Part Segmentation arXiv:2204.04655