# Audio-Visual Segmentation on AVSBench S4

Source: [Audio-Visual Segmentation on AVSBench S4](<https://www.sota2.com/research/sota/audio-visual-segmentation-on-avsbench-s4>)

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

Data updated: 2026-05-12T09:51:10.844Z

Evaluation of Audio-Visual Segmentation performance on the AVSBench S4 dataset, typically reporting 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 | Audio-Visual Segmentation |
| Dataset (including any stated split) | AVSBench S4 |
| Metric | MJ Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 12 of 12 |
| 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

- [MJ Score](<https://www.sota2.com/research/sota/audio-visual-segmentation-on-avsbench-s4.md?metric=MJ%20Score>): Higher is better; unit: PERCENT.

- [mIoU](<https://www.sota2.com/research/sota/audio-visual-segmentation-on-avsbench-s4.md?metric=mIoU>): Higher is better; unit: PERCENT.

- [MF Score](<https://www.sota2.com/research/sota/audio-visual-segmentation-on-avsbench-s4.md?metric=MF%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | MJ Score | Result date | Report paper |
| --- | --- | --- | --- |
| SelM | 76.6 | Not recorded | 6a02b403dbc836b87d8245ce |
| AVSegFormer | 76.5 | Not recorded | 6a02b403dbc836b87d8245ce |
| LightAVSeg | 75.6 | Not recorded | 6a02b403dbc836b87d8245ce |
| AVSBench | 72.8 | Not recorded | 6a02b403dbc836b87d8245ce |
| SelM | 59.1 | Not recorded | 6a02b403dbc836b87d8245ce |
| AVSegFormer | 53.8 | Not recorded | 6a02b403dbc836b87d8245ce |
| AVSBench | 47.9 | Not recorded | 6a02b403dbc836b87d8245ce |

### SelM

MJ Score: 76.6. Result ID: 6a02f80e47f34ce924e62dc9.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b403dbc836b87d8245ce.md>)

Recorded evaluation configuration: {"Visual Backbone":"R50","Audio Backbone":"VGGish","GPU Latency (ms)":"25.3","Mobile Latency (ms)":"1003.8","Params":"117.6M"}. This record does not assert independent reproduction.

### AVSegFormer

MJ Score: 76.5. Result ID: 6a02f80e47f34ce924e62dc8.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b403dbc836b87d8245ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5174e.md>)

Recorded evaluation configuration: {"Visual Backbone":"R50","Audio Backbone":"VGGish","GPU Latency (ms)":"29.0","Mobile Latency (ms)":"1271.4","Params":"151.1M"}. This record does not assert independent reproduction.

### LightAVSeg

MJ Score: 75.6. Result ID: 6a02f80e47f34ce924e62dcd.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b403dbc836b87d8245ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a02b403dbc836b87d8245ce.md>)

Recorded evaluation configuration: {"Visual Backbone":"Sea","Audio Backbone":"MNetV2","GPU Latency (ms)":"15.9","Mobile Latency (ms)":"163.4","Params":"20.5M"}. This record does not assert independent reproduction.

### AVSBench

MJ Score: 72.8. Result ID: 6a02f80e47f34ce924e62dc7.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b403dbc836b87d8245ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e8e1e209131cea408f0.md>)

Recorded evaluation configuration: {"Visual Backbone":"R50","Audio Backbone":"VGGish","GPU Latency (ms)":"21.2","Mobile Latency (ms)":"753.5","Params":"91.4M"}. This record does not assert independent reproduction.

### SelM

MJ Score: 59.1. Result ID: 6a02f80e47f34ce924e62dcc.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b403dbc836b87d8245ce.md>)

Recorded evaluation configuration: {"Visual Backbone":"Sea","Audio Backbone":"MNetV2","GPU Latency (ms)":"18.7","Mobile Latency (ms)":"308.6","Params":"39.5M"}. This record does not assert independent reproduction.

### AVSegFormer

MJ Score: 53.8. Result ID: 6a02f80e47f34ce924e62dcb.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b403dbc836b87d8245ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71791e209131cea5174e.md>)

Recorded evaluation configuration: {"Visual Backbone":"Sea","Audio Backbone":"MNetV2","GPU Latency (ms)":"22.2","Mobile Latency (ms)":"432.6","Params":"51.0M"}. This record does not assert independent reproduction.

### AVSBench

MJ Score: 47.9. Result ID: 6a02f80e47f34ce924e62dca.

[Reporting paper](<https://www.sota2.com/research/paper/6a02b403dbc836b87d8245ce.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e8e1e209131cea408f0.md>)

Recorded evaluation configuration: {"Visual Backbone":"Sea","Audio Backbone":"MNetV2","GPU Latency (ms)":"19.5","Mobile Latency (ms)":"237.1","Params":"30.2M"}. This record does not assert independent reproduction.

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