Referring Expression Segmentation
The task aims at labeling the pixels of an image or video that represent an object instance referred by a linguistic expression. In particular, the referring expression (RE) must allow the identification of an individual object in a discourse or scene (the referent). REs unambiguously identify the target instance.
Papers
Showing 1–10 of 145 papers
All datasetsRefCoCo valRefCOCO testARefer-YouTube-VOS (2021 public validation)RefCOCO+ test BA2D SentencesRefCOCOg-valJ-HMDBDAVIS 2017 (val)RefCOCOg-testRefCOCO testBPhraseCutRefCOCO
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | DeRIS-L | Mean IoU | 78.59 | — | Unverified |
| 2 | MLCD-Seg-7B | Overall IoU | 75.6 | — | Unverified |
| 3 | HyperSeg | Overall IoU | 75.2 | — | Unverified |
| 4 | EVF-SAM | Overall IoU | 71.9 | — | Unverified |
| 5 | DETRIS | Overall IoU | 70.2 | — | Unverified |
| 6 | C3VG | Overall IoU | 68.95 | — | Unverified |
| 7 | UniLSeg-100 | Overall IoU | 68.15 | — | Unverified |
| 8 | UniLSeg-20 | Overall IoU | 66.99 | — | Unverified |
| 9 | UNINEXT-H | Overall IoU | 66.22 | — | Unverified |
| 10 | GROUNDHOG | Overall IoU | 64.9 | — | Unverified |