# Mask-guided defect generation on Anomaly-56K (test)

Source: [Mask-guided defect generation on Anomaly-56K (test)](<https://www.sota2.com/research/sota/mask-guided-defect-generation-on-anomaly-56k-test>)

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

Data updated: 2026-04-15T08:52:33.687Z

Evaluation of generative model performance on the test split of the Anomaly-56K dataset, focusing on defect generation quality measured by PSNR (Full, BG, Mask) and LPIPS (Mask).

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 | Mask-guided defect generation |
| Dataset (including any stated split) | Anomaly-56K (test) |
| Metric | PSNR (Full) |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [PSNR (Full)](<https://www.sota2.com/research/sota/mask-guided-defect-generation-on-anomaly-56k-test.md?metric=PSNR%20(Full)>): Higher is better; unit: SCALAR.

- [PSNR (BG)](<https://www.sota2.com/research/sota/mask-guided-defect-generation-on-anomaly-56k-test.md?metric=PSNR%20(BG)>): Higher is better; unit: SCALAR.

- [PSNR (Mask)](<https://www.sota2.com/research/sota/mask-guided-defect-generation-on-anomaly-56k-test.md?metric=PSNR%20(Mask)>): Higher is better; unit: SCALAR.

- [LPIPS (Mask)](<https://www.sota2.com/research/sota/mask-guided-defect-generation-on-anomaly-56k-test.md?metric=LPIPS%20(Mask)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | PSNR (Full) | Result date | Report paper |
| --- | --- | --- | --- |
| IAD-Unify | 30.03 | Not recorded | 69df1b7a7d0172140ec7b83b |
| Generation-only | 29.45 | Not recorded | 69df1b7a7d0172140ec7b83b |
| SD2 Inpainting | 28.92 | Not recorded | 69df1b7a7d0172140ec7b83b |

### IAD-Unify

PSNR (Full): 30.03. Result ID: 69df50d247f34ce924c82d0e.

[Reporting paper](<https://www.sota2.com/research/paper/69df1b7a7d0172140ec7b83b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69df1b7a7d0172140ec7b83b.md>)

Recorded evaluation configuration: {"Region expert":"true","Matched-pair protocol":"true"}. This record does not assert independent reproduction.

### Generation-only

PSNR (Full): 29.45. Result ID: 69df50d247f34ce924c82d0f.

[Reporting paper](<https://www.sota2.com/research/paper/69df1b7a7d0172140ec7b83b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69df1b7a7d0172140ec7b83b.md>)

Recorded evaluation configuration: {"Backbone":"Same as IAD-Unify","Region expert":"false","Matched-pair protocol":"true"}. This record does not assert independent reproduction.

### SD2 Inpainting

PSNR (Full): 28.92. Result ID: 69df50d247f34ce924c82d10.

[Reporting paper](<https://www.sota2.com/research/paper/69df1b7a7d0172140ec7b83b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c391e209131cea32e4a.md>)

[Implementation](<https://github.com/CompVis/latent-diffusion>)

Recorded evaluation configuration: {"Matched-pair protocol":"true"}. This record does not assert independent reproduction.

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