Panoptic Segmentation
Panoptic Segmentation is a computer vision task that combines semantic segmentation and instance segmentation to provide a comprehensive understanding of the scene. The goal of panoptic segmentation is to segment the image into semantically meaningful parts or regions, while also detecting and distinguishing individual instances of objects within those regions. In a given image, every pixel is assigned a semantic label, and pixels belonging to "things" classes (countable objects with instances, like cars and people) are assigned unique instance IDs. ( Image credit: Detectron2 )
Papers
Showing 1–10 of 462 papers
All datasetsCOCO test-devCityscapes valCOCO minivalADE20K valMapillary valCityscapes testLaRSS3DIS Area5ScanNetV2Indian Driving DatasetKITTI Panoptic SegmentationPanNuke
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Mask DINO (single scale) | PQ | 59.5 | — | Unverified |
| 2 | kMaX-DeepLab (single-scale) | PQ | 58.5 | — | Unverified |
| 3 | Mask2Former (Swin-L) | PQ | 58.3 | — | Unverified |
| 4 | Panoptic SegFormer (Swin-L) | PQ | 56.2 | — | Unverified |
| 5 | Panoptic SegFormer (PVTv2-B5) | PQ | 55.8 | — | Unverified |
| 6 | CMT-DeepLab (single-scale) | PQ | 55.7 | — | Unverified |
| 7 | K-Net (Swin-L) | PQ | 55.2 | — | Unverified |
| 8 | MaskConver (ResNet50, single-scale) | PQ | 53.6 | — | Unverified |
| 9 | MaskFormer (Swin-L) | PQ | 53.3 | — | Unverified |
| 10 | Panoptic FCN* (Swin-L) | PQ | 52.7 | — | Unverified |