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 | HINet | IoU overall | 0.65 | — | Unverified |
| 2 | ClawCraneNet | IoU overall | 0.64 | — | Unverified |
| 3 | CMSA+CFSA | IoU overall | 0.63 | — | Unverified |
| 4 | RefVOS | IoU overall | 0.61 | — | Unverified |
| 5 | SgMg (Video-Swin-B) | AP | 0.45 | — | Unverified |
| 6 | SOC (Video-Swin-B) | AP | 0.45 | — | Unverified |
| 7 | VLIDE | AP | 0.44 | — | Unverified |
| 8 | SOC (Video-Swin-T) | AP | 0.4 | — | Unverified |
| 9 | MTTR (w=10) | AP | 0.39 | — | Unverified |
| 10 | MTTR (w=8) | AP | 0.37 | — | Unverified |