# Referring Video Object Segmentation on A2D Sentences v1.0 (test)

Source: [Referring Video Object Segmentation on A2D Sentences v1.0 (test)](<https://www.sota2.com/research/sota/referring-video-object-segmentation-on-a2d-sentences-v1-0-test>)

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

Data updated: 2026-02-26T05:47:29.545Z

Evaluation of referring video object segmentation performance on the test split of the A2D Sentences v1.0 dataset, measured using precision thresholds (P@0.5 to P@0.9), mAP, Overall IoU, and Mean IoU.

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 | Referring Video Object Segmentation |
| Dataset (including any stated split) | A2D Sentences v1.0 (test) |
| Metric | IoU Overall |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 26 of 26 |
| 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

- [IoU Overall](<https://www.sota2.com/research/sota/referring-video-object-segmentation-on-a2d-sentences-v1-0-test.md?metric=IoU%20Overall>): Higher is better; unit: PERCENT.

- [P@0.5](<https://www.sota2.com/research/sota/referring-video-object-segmentation-on-a2d-sentences-v1-0-test.md?metric=P%400.5>): Higher is better; unit: PERCENT.

- [P@0.6](<https://www.sota2.com/research/sota/referring-video-object-segmentation-on-a2d-sentences-v1-0-test.md?metric=P%400.6>): Higher is better; unit: PERCENT.

- [P@0.7](<https://www.sota2.com/research/sota/referring-video-object-segmentation-on-a2d-sentences-v1-0-test.md?metric=P%400.7>): Higher is better; unit: PERCENT.

- [P@0.8](<https://www.sota2.com/research/sota/referring-video-object-segmentation-on-a2d-sentences-v1-0-test.md?metric=P%400.8>): Higher is better; unit: PERCENT.

- [P@0.9](<https://www.sota2.com/research/sota/referring-video-object-segmentation-on-a2d-sentences-v1-0-test.md?metric=P%400.9>): Higher is better; unit: PERCENT.

- [mAP](<https://www.sota2.com/research/sota/referring-video-object-segmentation-on-a2d-sentences-v1-0-test.md?metric=mAP>): Higher is better; unit: PERCENT.

- [IoU Mean](<https://www.sota2.com/research/sota/referring-video-object-segmentation-on-a2d-sentences-v1-0-test.md?metric=IoU%20Mean>): Higher is better; unit: PERCENT.

## Recorded results

| Method | IoU Overall | Result date | Report paper |
| --- | --- | --- | --- |
| LoSh-S | 81.2 | Not recorded | 68fa72c01e209131cea58623 |
| SgMg | 79.9 | Not recorded | 68fa72c01e209131cea58623 |
| LoSh-S | 79.3 | Not recorded | 68fa72c01e209131cea58623 |
| ReferFormer | 78.6 | Not recorded | 68fa72c01e209131cea58623 |
| SgMg | 78 | Not recorded | 68fa72c01e209131cea58623 |
| LoSh-M | 72.9 | Not recorded | 68fa72c01e209131cea58623 |
| LoSh-M w/ MSh | 72.9 | Not recorded | 68fa72c01e209131cea58623 |
| LoSh-M w/ Bi-CA | 72.8 | Not recorded | 68fa72c01e209131cea58623 |
| MANet | 72.6 | Not recorded | 68fa6d1e1e209131cea37e48 |
| LoSh-M w/o CA | 72.5 | Not recorded | 68fa72c01e209131cea58623 |
| LoSh-M w/ Inv-CA | 72.5 | Not recorded | 68fa72c01e209131cea58623 |
| LoSh-M w/o Llsi | 72.3 | Not recorded | 68fa72c01e209131cea58623 |
| LoSh-M w/o Sh | 71.1 | Not recorded | 68fa72c01e209131cea58623 |
| MTTR | 70.2 | Not recorded | 68fa6d1e1e209131cea37e48 |
| MTTR | 70.2 | Not recorded | 68fa72c01e209131cea58623 |
| CSTM | 66.2 | Not recorded | 68fa6d1e1e209131cea37e48 |
| PolarRPE | 66.1 | Not recorded | 68fa6d1e1e209131cea37e48 |
| CMPC-V | 65.3 | Not recorded | 68fa6d1e1e209131cea37e48 |
| CMPC-V | 65.3 | Not recorded | 68fa72c01e209131cea58623 |
| CCMA | 63.2 | Not recorded | 68fa6d1e1e209131cea37e48 |
| CMDy | 62.3 | Not recorded | 68fa6d1e1e209131cea37e48 |
| CMSA+CFSA | 61.8 | Not recorded | 68fa6d1e1e209131cea37e48 |
| ACGA | 60.1 | Not recorded | 68fa6d1e1e209131cea37e48 |
| VT-Capsule | 56.8 | Not recorded | 68fa6d1e1e209131cea37e48 |
| A2DS | 55.1 | Not recorded | 68fa6d1e1e209131cea37e48 |
| Gavrilyuk et al. | 53.6 | Not recorded | 68fa72c01e209131cea58623 |

### LoSh-S

IoU Overall: 81.2. Result ID: 699291963a6c3148490fda29.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-B","Number of input frames (w)":5}. This record does not assert independent reproduction.

### SgMg

IoU Overall: 79.9. Result ID: 699291963a6c3148490fda28.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70261e209131cea496a0.md>)

[Implementation](<https://github.com/bo-miao/SgMg>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-B","Number of input frames (w)":5}. This record does not assert independent reproduction.

### LoSh-S

IoU Overall: 79.3. Result ID: 699291963a6c3148490fda27.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-T","Number of input frames (w)":5}. This record does not assert independent reproduction.

### ReferFormer

IoU Overall: 78.6. Result ID: 699291963a6c3148490fda23.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bff1e209131cea31875.md>)

[Implementation](<https://github.com/wjn922/ReferFormer>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-B"}. This record does not assert independent reproduction.

### SgMg

IoU Overall: 78. Result ID: 699291963a6c3148490fda26.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70261e209131cea496a0.md>)

[Implementation](<https://github.com/bo-miao/SgMg>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-T","Number of input frames (w)":5}. This record does not assert independent reproduction.

### LoSh-M

IoU Overall: 72.9. Result ID: 699291963a6c3148490fda25.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-T","Number of input frames (w)":8}. This record does not assert independent reproduction.

### LoSh-M w/ MSh

IoU Overall: 72.9. Result ID: 699291963a6c3148490fda2f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-T"}. This record does not assert independent reproduction.

### LoSh-M w/ Bi-CA

IoU Overall: 72.8. Result ID: 699291963a6c3148490fda2e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-T"}. This record does not assert independent reproduction.

### MANet

IoU Overall: 72.6. Result ID: 698f932f3a6c314849fa3a2a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

[Implementation](<https://github.com/DexiangHong/MANet>)

Recorded evaluation configuration: {"Venue":"-"}. This record does not assert independent reproduction.

### LoSh-M w/o CA

IoU Overall: 72.5. Result ID: 699291963a6c3148490fda2c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-T"}. This record does not assert independent reproduction.

### LoSh-M w/ Inv-CA

IoU Overall: 72.5. Result ID: 699291963a6c3148490fda2d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-T"}. This record does not assert independent reproduction.

### LoSh-M w/o Llsi

IoU Overall: 72.3. Result ID: 699291963a6c3148490fda2b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-T"}. This record does not assert independent reproduction.

### LoSh-M w/o Sh

IoU Overall: 71.1. Result ID: 699291963a6c3148490fda2a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-T"}. This record does not assert independent reproduction.

### MTTR

IoU Overall: 70.2. Result ID: 698f932f3a6c314849fa3a29.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32c7c.md>)

[Implementation](<https://github.com/mttr2021/MTTR>)

Recorded evaluation configuration: {"Venue":"CVPR22"}. This record does not assert independent reproduction.

### MTTR

IoU Overall: 70.2. Result ID: 699291963a6c3148490fda24.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c341e209131cea32c7c.md>)

[Implementation](<https://github.com/mttr2021/MTTR>)

Recorded evaluation configuration: {"Backbone":"Video-Swin-T","Number of input frames (w)":8}. This record does not assert independent reproduction.

### CSTM

IoU Overall: 66.2. Result ID: 698f932f3a6c314849fa3a25.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69f31e209131cea25308.md>)

Recorded evaluation configuration: {"Venue":"CVPR21"}. This record does not assert independent reproduction.

### PolarRPE

IoU Overall: 66.1. Result ID: 698f932f3a6c314849fa3a24.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

Recorded evaluation configuration: {"Venue":"IJCAI20"}. This record does not assert independent reproduction.

### CMPC-V

IoU Overall: 65.3. Result ID: 698f932f3a6c314849fa3a26.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69f41e209131cea253e9.md>)

[Implementation](<https://github.com/spyflying/CMPC-Refseg>)

Recorded evaluation configuration: {"Venue":"T-PAMI21"}. This record does not assert independent reproduction.

### CMPC-V

IoU Overall: 65.3. Result ID: 699291963a6c3148490fda22.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69f41e209131cea253e9.md>)

[Implementation](<https://github.com/spyflying/CMPC-Refseg>)

Recorded evaluation configuration: {"Backbone":"I3D"}. This record does not assert independent reproduction.

### CCMA

IoU Overall: 63.2. Result ID: 698f932f3a6c314849fa3a28.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

Recorded evaluation configuration: {"Venue":"ACMMM21"}. This record does not assert independent reproduction.

### CMDy

IoU Overall: 62.3. Result ID: 698f932f3a6c314849fa3a22.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

Recorded evaluation configuration: {"Venue":"AAAI20"}. This record does not assert independent reproduction.

### CMSA+CFSA

IoU Overall: 61.8. Result ID: 698f932f3a6c314849fa3a27.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69371e209131cea20e21.md>)

Recorded evaluation configuration: {"Venue":"T-PAMI21"}. This record does not assert independent reproduction.

### ACGA

IoU Overall: 60.1. Result ID: 698f932f3a6c314849fa3a21.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

Recorded evaluation configuration: {"Venue":"ICCV19"}. This record does not assert independent reproduction.

### VT-Capsule

IoU Overall: 56.8. Result ID: 698f932f3a6c314849fa3a23.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

Recorded evaluation configuration: {"Venue":"CVPR20"}. This record does not assert independent reproduction.

### A2DS

IoU Overall: 55.1. Result ID: 698f932f3a6c314849fa3a20.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d1e1e209131cea37e48.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64261e209131cea02e24.md>)

[Implementation](<https://kgavrilyuk.github.io/publication/actor_action/>)

Recorded evaluation configuration: {"Venue":"CVPR18"}. This record does not assert independent reproduction.

### Gavrilyuk et al.

IoU Overall: 53.6. Result ID: 699291963a6c3148490fda21.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72c01e209131cea58623.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64261e209131cea02e24.md>)

[Implementation](<https://kgavrilyuk.github.io/publication/actor_action/>)

Recorded evaluation configuration: {"Backbone":"I3D"}. This record does not assert independent reproduction.

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