# Visual Anomaly Segmentation on RSDD (AP, F1-Max, IoUano)

Source: [Visual Anomaly Segmentation on RSDD (AP, F1-Max, IoUano)](<https://www.sota2.com/research/sota/visual-anomaly-segmentation-on-rsdd-ap-f1-max-iouano>)

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

Data updated: 2026-03-04T13:42:51.750Z

Evaluation of Visual Anomaly Segmentation performance on the RSDD dataset, reporting AP, F1-Max, and IoUano scores.

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 | Visual Anomaly Segmentation |
| Dataset (including any stated split) | RSDD |
| Metric | AP |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 15 of 15 |
| 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

- [AP](<https://www.sota2.com/research/sota/visual-anomaly-segmentation-on-rsdd-ap-f1-max-iouano.md?metric=AP>): Higher is better; unit: PERCENT.

- [F1-Max](<https://www.sota2.com/research/sota/visual-anomaly-segmentation-on-rsdd-ap-f1-max-iouano.md?metric=F1-Max>): Higher is better; unit: PERCENT.

- [IoUano](<https://www.sota2.com/research/sota/visual-anomaly-segmentation-on-rsdd-ap-f1-max-iouano.md?metric=IoUano>): Higher is better; unit: PERCENT.

## Recorded results

| Method | AP | Result date | Report paper |
| --- | --- | --- | --- |
| AG-VAS | 56.2 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| AG-VAS | 54.3 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| LISA\* | 52.1 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| LISA\* | 48.8 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| Bayes-PFL\* | 41.9 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| Bayes-PFL | 39.1 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| LISA | 31.6 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| APRIL-GAN | 30.6 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| AnomalyCLIP\* | 25.5 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| AnomalyCLIP | 19.1 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| LISA | 6.9 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| PixelLM | 5.3 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| LISA | 5.2 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| WinCLIP | 2.1 | Not recorded | 69a66c1a2d603e6de3e3ce76 |
| PaDT | 0.9 | Not recorded | 69a66c1a2d603e6de3e3ce76 |

### AG-VAS

AP: 56.2. Result ID: 69a836c06dbb7075b94ec3f8.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

Recorded evaluation configuration: {"Base Model":"LLaVA-OneVision-7B"}. This record does not assert independent reproduction.

### AG-VAS

AP: 54.3. Result ID: 69a836c06dbb7075b94ec3f7.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

Recorded evaluation configuration: {"Base Model":"LLaVA-13B"}. This record does not assert independent reproduction.

### LISA\*

AP: 52.1. Result ID: 69a836c06dbb7075b94ec3f4.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa730a1e209131cea5a0b2.md>)

[Implementation](<https://www.sota2.com/github.com/dvlab-research/LISA>)

Recorded evaluation configuration: {"Base Model":"LLaVA-13B","Retrained":"true"}. This record does not assert independent reproduction.

### LISA\*

AP: 48.8. Result ID: 69a836c06dbb7075b94ec3f6.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa730a1e209131cea5a0b2.md>)

[Implementation](<https://www.sota2.com/github.com/dvlab-research/LISA>)

Recorded evaluation configuration: {"Base Model":"LLaVA-OneVision-7B","Retrained":"true"}. This record does not assert independent reproduction.

### Bayes-PFL\*

AP: 41.9. Result ID: 69a836c06dbb7075b94ec3ef.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79151e209131cea7c094.md>)

[Implementation](<https://github.com/xiaozhen228/Bayes-PFL>)

Recorded evaluation configuration: {"Base Model":"CLIP(ViT-L-14-336)","Retrained":"true"}. This record does not assert independent reproduction.

### Bayes-PFL

AP: 39.1. Result ID: 69a836c06dbb7075b94ec3ee.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79151e209131cea7c094.md>)

[Implementation](<https://github.com/xiaozhen228/Bayes-PFL>)

Recorded evaluation configuration: {"Base Model":"CLIP(ViT-L-14-336)"}. This record does not assert independent reproduction.

### LISA

AP: 31.6. Result ID: 69a836c06dbb7075b94ec3f5.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa730a1e209131cea5a0b2.md>)

[Implementation](<https://www.sota2.com/github.com/dvlab-research/LISA>)

Recorded evaluation configuration: {"Base Model":"LLaVA-OneVision-7B"}. This record does not assert independent reproduction.

### APRIL-GAN

AP: 30.6. Result ID: 69a836c06dbb7075b94ec3eb.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70d71e209131cea4d95d.md>)

[Implementation](<https://github.com/ByChelsea/VAND-APRIL-GAN>)

Recorded evaluation configuration: {"Base Model":"CLIP(ViT-L-14-336)"}. This record does not assert independent reproduction.

### AnomalyCLIP\*

AP: 25.5. Result ID: 69a836c06dbb7075b94ec3ed.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78451e209131cea77851.md>)

[Implementation](<https://github.com/zqhang/AnomalyCLIP>)

Recorded evaluation configuration: {"Base Model":"CLIP(ViT-L-14-336)","Retrained":"true"}. This record does not assert independent reproduction.

### AnomalyCLIP

AP: 19.1. Result ID: 69a836c06dbb7075b94ec3ec.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78451e209131cea77851.md>)

[Implementation](<https://github.com/zqhang/AnomalyCLIP>)

Recorded evaluation configuration: {"Base Model":"CLIP(ViT-L-14-336)"}. This record does not assert independent reproduction.

### LISA

AP: 6.9. Result ID: 69a836c06dbb7075b94ec3f2.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa730a1e209131cea5a0b2.md>)

[Implementation](<https://www.sota2.com/github.com/dvlab-research/LISA>)

Recorded evaluation configuration: {"Base Model":"LLaVA-7B"}. This record does not assert independent reproduction.

### PixelLM

AP: 5.3. Result ID: 69a836c06dbb7075b94ec3f0.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74241e209131cea5ff3b.md>)

[Implementation](<https://pixellm.github.io/>)

Recorded evaluation configuration: {"Base Model":"LLaVA-13B"}. This record does not assert independent reproduction.

### LISA

AP: 5.2. Result ID: 69a836c06dbb7075b94ec3f3.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa730a1e209131cea5a0b2.md>)

[Implementation](<https://www.sota2.com/github.com/dvlab-research/LISA>)

Recorded evaluation configuration: {"Base Model":"LLaVA-13B"}. This record does not assert independent reproduction.

### WinCLIP

AP: 2.1. Result ID: 69a836c06dbb7075b94ec3ea.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6eea1e209131cea42e2d.md>)

Recorded evaluation configuration: {"Base Model":"CLIP(ViT-B-16-240)"}. This record does not assert independent reproduction.

### PaDT

AP: 0.9. Result ID: 69a836c06dbb7075b94ec3f1.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c1a2d603e6de3e3ce76.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bd51e209131cea8ac9c.md>)

[Implementation](<https://github.com/Gorilla-Lab-SCUT/PaDT>)

Recorded evaluation configuration: {"Base Model":"Qwen2.5-VL-7B"}. This record does not assert independent reproduction.

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