# Open-Vocabulary Semantic Segmentation on VOC 21

Source: [Open-Vocabulary Semantic Segmentation on VOC 21](<https://www.sota2.com/research/sota/open-vocabulary-semantic-segmentation-on-voc-21>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-10T09:40:49.515Z

Evaluation of Open-Vocabulary Semantic Segmentation performance on the VOC 21 dataset, reporting Mean Intersection over Union (mIoU).

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Open-Vocabulary Semantic Segmentation |
| Dataset (including any stated split) | VOC 21 |
| Metric | mIoU |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 9 of 9 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [mIoU](<https://www.sota2.com/research/sota/open-vocabulary-semantic-segmentation-on-voc-21.md?metric=mIoU>): Higher is better; unit: PERCENT.

## Recorded results

| Method | mIoU | Result date | Report paper |
| --- | --- | --- | --- |
| DSGS | 67.7 | Not recorded | 69d884407d0172140ec77c62 |
| OVS-DINO | 64 | Not recorded | 69d884407d0172140ec77c62 |
| SynSeg | 62.2 | Not recorded | 69d884407d0172140ec77c62 |
| CLIP-DINOiser | 62.1 | Not recorded | 69d884407d0172140ec77c62 |
| Talk2DINO | 61.5 | Not recorded | 69d884407d0172140ec77c62 |
| TCL | 51.2 | Not recorded | 69d884407d0172140ec77c62 |
| GroupViT | 50.4 | Not recorded | 69d884407d0172140ec77c62 |
| CoDe | 46 | Not recorded | 69d884407d0172140ec77c62 |
| ReCo | 25.1 | Not recorded | 69d884407d0172140ec77c62 |

### DSGS

mIoU: 67.7. Result ID: 69d8c59747f34ce924c21060.

[Reporting paper](<https://www.sota2.com/research/paper/69d884407d0172140ec77c62.md>)

Recorded evaluation configuration: {"Backbone":"CLIP B/16"}. This record does not assert independent reproduction.

### OVS-DINO

mIoU: 64. Result ID: 69d8c59747f34ce924c21061.

[Reporting paper](<https://www.sota2.com/research/paper/69d884407d0172140ec77c62.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d884407d0172140ec77c62.md>)

Recorded evaluation configuration: {"Backbone":"DINOv2 B/14"}. This record does not assert independent reproduction.

### SynSeg

mIoU: 62.2. Result ID: 69d8c59747f34ce924c2105e.

[Reporting paper](<https://www.sota2.com/research/paper/69d884407d0172140ec77c62.md>)

Recorded evaluation configuration: {"Backbone":"CLIP B/16"}. This record does not assert independent reproduction.

### CLIP-DINOiser

mIoU: 62.1. Result ID: 69d8c59747f34ce924c2105d.

[Reporting paper](<https://www.sota2.com/research/paper/69d884407d0172140ec77c62.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72b51e209131cea58110.md>)

Recorded evaluation configuration: {"Backbone":"CLIP B/16"}. This record does not assert independent reproduction.

### Talk2DINO

mIoU: 61.5. Result ID: 69d8c59747f34ce924c2105f.

[Reporting paper](<https://www.sota2.com/research/paper/69d884407d0172140ec77c62.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b4e1e209131cea87a60.md>)

Recorded evaluation configuration: {"Backbone":"DINOv2 B/14"}. This record does not assert independent reproduction.

### TCL

mIoU: 51.2. Result ID: 69d8c59747f34ce924c2105b.

[Reporting paper](<https://www.sota2.com/research/paper/69d884407d0172140ec77c62.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ee81e209131cea42ce9.md>)

[Implementation](<https://github.com/kakaobrain/tcl>)

Recorded evaluation configuration: {"Backbone":"CLIP B/16"}. This record does not assert independent reproduction.

### GroupViT

mIoU: 50.4. Result ID: 69d8c59747f34ce924c21059.

[Reporting paper](<https://www.sota2.com/research/paper/69d884407d0172140ec77c62.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cfc1e209131cea371f1.md>)

[Implementation](<https://github.com/NVlabs/GroupViT>)

Recorded evaluation configuration: {"Backbone":"CLIP S/16"}. This record does not assert independent reproduction.

### CoDe

mIoU: 46. Result ID: 69d8c59747f34ce924c2105c.

[Reporting paper](<https://www.sota2.com/research/paper/69d884407d0172140ec77c62.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72d51e209131cea58cce.md>)

[Implementation](<https://github.com/072jiajia/image-text-co-decomposition>)

Recorded evaluation configuration: {"Backbone":"CLIP B/16"}. This record does not assert independent reproduction.

### ReCo

mIoU: 25.1. Result ID: 69d8c59747f34ce924c2105a.

[Reporting paper](<https://www.sota2.com/research/paper/69d884407d0172140ec77c62.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cc81e209131cea35f62.md>)

[Implementation](<https://www.robots.ox.ac.uk/%E1%B9%BDgg/research/reco.%20(Project%20page%20provided,%20not%20strict%20code%20repository)>)

Recorded evaluation configuration: {"Backbone":"CLIP B/16"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.