Instance Segmentation
Instance Segmentation is a computer vision task that involves identifying and separating individual objects within an image, including detecting the boundaries of each object and assigning a unique label to each object. The goal of instance segmentation is to produce a pixel-wise segmentation map of the image, where each pixel is assigned to a specific object instance.
Image Credit: Deep Occlusion-Aware Instance Segmentation with Overlapping BiLayers, CVPR'21
Papers
Showing 1–10 of 2262 papers
All datasetsCOCO test-devCOCO minivalLVIS v1.0 valCityscapes valADE20K valARMBenchOccluded COCOOoDISSeparated COCOTBBRBDD100K valCOCO val (panoptic labels)
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | InternImage-B | GFLOPs | 501 | — | Unverified |
| 2 | Co-DETR | mask AP | 56.6 | — | Unverified |
| 3 | ViT-CoMer-L (Mask RCNN, DINOv2) | mask AP | 55.9 | — | Unverified |
| 4 | InternImage-H | mask AP | 55.4 | — | Unverified |
| 5 | EVA | mask AP | 55 | — | Unverified |
| 6 | Mask Frozen-DETR | mask AP | 54.9 | — | Unverified |
| 7 | MasK DINO (SwinL, multi-scale) | mask AP | 54.5 | — | Unverified |
| 8 | ViT-Adapter-L (HTC++, BEiTv2, O365, multi-scale) | mask AP | 54.2 | — | Unverified |
| 9 | GLEE-Pro | mask AP | 54.2 | — | Unverified |
| 10 | SwinV2-G (HTC++) | mask AP | 53.7 | — | Unverified |