# Image Restoration on Screen-Compound MRL (test)

Source: [Image Restoration on Screen-Compound MRL (test)](<https://www.sota2.com/research/sota/image-restoration-on-screen-compound-mrl-test>)

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

Data updated: 2026-03-09T06:59:19.920Z

Evaluation of Image Restoration performance on the test split of the Screen-Compound MRL dataset, measured using PSNR, SSIM, and LPIPS.

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 | Image Restoration |
| Dataset (including any stated split) | Screen-Compound MRL (test) |
| Metric | PSNR |
| Unit | SCALAR |
| 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

- [PSNR](<https://www.sota2.com/research/sota/image-restoration-on-screen-compound-mrl-test.md?metric=PSNR>): Higher is better; unit: SCALAR.

- [SSIM](<https://www.sota2.com/research/sota/image-restoration-on-screen-compound-mrl-test.md?metric=SSIM>): Higher is better; unit: SCALAR.

- [LPIPS](<https://www.sota2.com/research/sota/image-restoration-on-screen-compound-mrl-test.md?metric=LPIPS>): Lower is better; unit: SCALAR.

## Recorded results

| Method | PSNR | Result date | Report paper |
| --- | --- | --- | --- |
| DeVeiler | 21.57 | Not recorded | 6923e675d6b0dd334292eed7 |
| SwinIR + Flare7K++ | 20.74 | Not recorded | 6923e675d6b0dd334292eed7 |
| QDMR | 20.67 | Not recorded | 6923e675d6b0dd334292eed7 |
| SwinIR | 19.34 | Not recorded | 6923e675d6b0dd334292eed7 |
| NAFNET | 18.91 | Not recorded | 6923e675d6b0dd334292eed7 |
| DiffBIR | 18.7 | Not recorded | 6923e675d6b0dd334292eed7 |
| UCL-Dehaze | 18.69 | Not recorded | 6923e675d6b0dd334292eed7 |
| SwinIR + RIDCP | 18.66 | Not recorded | 6923e675d6b0dd334292eed7 |
| DiffDehaze | 18.55 | Not recorded | 6923e675d6b0dd334292eed7 |
| SwinIR + DiffDehaze | 18.28 | Not recorded | 6923e675d6b0dd334292eed7 |
| CycleGAN | 17.8 | Not recorded | 6923e675d6b0dd334292eed7 |
| Fast two-step | 14.94 | Not recorded | 6923e675d6b0dd334292eed7 |

### DeVeiler

PSNR: 21.57. Result ID: 69ae6f286dbb7075b955fb9a.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

Recorded evaluation configuration: {"Pipeline Type":"Domain Adaptation Pipeline","Latency(s)":"0.387"}. This record does not assert independent reproduction.

### SwinIR + Flare7K++

PSNR: 20.74. Result ID: 69ae6f286dbb7075b955fb95.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

Recorded evaluation configuration: {"Pipeline Type":"Cascaded Pipeline","Pre-trained":"true","Latency(s)":"0.990"}. This record does not assert independent reproduction.

### QDMR

PSNR: 20.67. Result ID: 69ae6f286dbb7075b955fb99.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa487c1e209131ce92ff57.md>)

Recorded evaluation configuration: {"Pipeline Type":"Domain Adaptation Pipeline","Latency(s)":"0.428"}. This record does not assert independent reproduction.

### SwinIR

PSNR: 19.34. Result ID: 69ae6f286dbb7075b955fb90.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a9a1e209131cea294aa.md>)

[Implementation](<https://github.com/JingyunLiang/SwinIR>)

Recorded evaluation configuration: {"Pipeline Type":"Single Degradation Pipeline","Latency(s)":"0.782"}. This record does not assert independent reproduction.

### NAFNET

PSNR: 18.91. Result ID: 69ae6f286dbb7075b955fb91.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d1d1e209131cea37def.md>)

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

Recorded evaluation configuration: {"Pipeline Type":"Single Degradation Pipeline","Latency(s)":"0.353"}. This record does not assert independent reproduction.

### DiffBIR

PSNR: 18.7. Result ID: 69ae6f286dbb7075b955fb92.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72e01e209131cea5918f.md>)

[Implementation](<https://github.com/XPixelGroup/DiffBIR>)

Recorded evaluation configuration: {"Pipeline Type":"Single Degradation Pipeline","Latency(s)":"66.13"}. This record does not assert independent reproduction.

### UCL-Dehaze

PSNR: 18.69. Result ID: 69ae6f286dbb7075b955fb97.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c7f1e209131cea345a8.md>)

Recorded evaluation configuration: {"Pipeline Type":"Domain Adaptation Pipeline","Latency(s)":"0.208"}. This record does not assert independent reproduction.

### SwinIR + RIDCP

PSNR: 18.66. Result ID: 69ae6f286dbb7075b955fb93.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

Recorded evaluation configuration: {"Pipeline Type":"Cascaded Pipeline","Pre-trained":"true","Latency(s)":"1.240"}. This record does not assert independent reproduction.

### DiffDehaze

PSNR: 18.55. Result ID: 69ae6f286dbb7075b955fb98.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77a41e209131cea74912.md>)

Recorded evaluation configuration: {"Pipeline Type":"Domain Adaptation Pipeline","Latency(s)":"91.76"}. This record does not assert independent reproduction.

### SwinIR + DiffDehaze

PSNR: 18.28. Result ID: 69ae6f286dbb7075b955fb94.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

Recorded evaluation configuration: {"Pipeline Type":"Cascaded Pipeline","Pre-trained":"true","Latency(s)":"92.54"}. This record does not assert independent reproduction.

### CycleGAN

PSNR: 17.8. Result ID: 69ae6f286dbb7075b955fb96.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68351e209131cea1ac7a.md>)

Recorded evaluation configuration: {"Pipeline Type":"Domain Adaptation Pipeline","Latency(s)":"0.208"}. This record does not assert independent reproduction.

### Fast two-step

PSNR: 14.94. Result ID: 69ae6f286dbb7075b955fb8f.

[Reporting paper](<https://www.sota2.com/research/paper/6923e675d6b0dd334292eed7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d1d1e209131cea37db3.md>)

Recorded evaluation configuration: {"Pipeline Type":"Single Degradation Pipeline","Latency(s)":"0.390"}. This record does not assert independent reproduction.

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