# Face Verification on DFW (test)

Source: [Face Verification on DFW (test)](<https://www.sota2.com/research/sota/face-verification-on-dfw-test>)

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

Data updated: 2026-02-26T05:45:40.167Z

Evaluation of face verification performance on the DFW test set, typically measured using GAR at specific FAR thresholds.

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 | Face Verification |
| Dataset (including any stated split) | DFW (test) |
| Metric | GAR @ 0.1% FAR |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [GAR @ 0.1% FAR](<https://www.sota2.com/research/sota/face-verification-on-dfw-test.md?metric=GAR%20%40%200.1%25%20FAR>): Higher is better; unit: PERCENT.

- [GAR @ 1% FAR](<https://www.sota2.com/research/sota/face-verification-on-dfw-test.md?metric=GAR%20%40%201%25%20FAR>): Higher is better; unit: PERCENT.

- [GAR @ 10% FAR](<https://www.sota2.com/research/sota/face-verification-on-dfw-test.md?metric=GAR%20%40%2010%25%20FAR>): Higher is better; unit: PERCENT.

## Recorded results

| Method | GAR @ 0.1% FAR | Result date | Report paper |
| --- | --- | --- | --- |
| DisguiseNet (LC + LR + LBCE) | 23.25 | Not recorded | 68fa644e1e209131cea03e63 |
| LC + LR | 21.32 | Not recorded | 68fa644e1e209131cea03e63 |
| LC | 21.03 | Not recorded | 68fa644e1e209131cea03e63 |
| LC + LR + LBCE(W) | 19.83 | Not recorded | 68fa644e1e209131cea03e63 |
| LC + LR(W) | 19.71 | Not recorded | 68fa644e1e209131cea03e63 |
| LC(W) | 19.54 | Not recorded | 68fa644e1e209131cea03e63 |
| Baseline | 17.73 | Not recorded | 68fa644e1e209131cea03e63 |

### DisguiseNet (LC + LR + LBCE)

GAR @ 0.1% FAR: 23.25. Result ID: 698c2bb3112aeb9460efecb6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Implementation](<https://github.com/pvskand/DisguiseNet>)

Recorded evaluation configuration: {"loss\_function":"Contrastive Loss (LC) + Regression Loss (LR) + Binary Cross-Entropy Loss (LBCE)","weakly\_labelled\_data":true}. This record does not assert independent reproduction.

### LC + LR

GAR @ 0.1% FAR: 21.32. Result ID: 698c2bb3112aeb9460efecb5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Implementation](<https://github.com/pvskand/DisguiseNet>)

Recorded evaluation configuration: {"loss\_function":"Contrastive Loss (LC) + Regression Loss (LR)","weakly\_labelled\_data":true}. This record does not assert independent reproduction.

### LC

GAR @ 0.1% FAR: 21.03. Result ID: 698c2bb3112aeb9460efecb4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Implementation](<https://github.com/pvskand/DisguiseNet>)

Recorded evaluation configuration: {"loss\_function":"Contrastive Loss (LC)","weakly\_labelled\_data":true}. This record does not assert independent reproduction.

### LC + LR + LBCE(W)

GAR @ 0.1% FAR: 19.83. Result ID: 698c2bb3112aeb9460efecb3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Implementation](<https://github.com/pvskand/DisguiseNet>)

Recorded evaluation configuration: {"loss\_function":"Contrastive Loss (LC) + Regression Loss (LR) + Binary Cross-Entropy Loss (LBCE)","weakly\_labelled\_data":false}. This record does not assert independent reproduction.

### LC + LR(W)

GAR @ 0.1% FAR: 19.71. Result ID: 698c2bb3112aeb9460efecb2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Implementation](<https://github.com/pvskand/DisguiseNet>)

Recorded evaluation configuration: {"loss\_function":"Contrastive Loss (LC) + Regression Loss (LR)","weakly\_labelled\_data":false}. This record does not assert independent reproduction.

### LC(W)

GAR @ 0.1% FAR: 19.54. Result ID: 698c2bb3112aeb9460efecb1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

[Implementation](<https://github.com/pvskand/DisguiseNet>)

Recorded evaluation configuration: {"loss\_function":"Contrastive Loss (LC)","weakly\_labelled\_data":false}. This record does not assert independent reproduction.

### Baseline

GAR @ 0.1% FAR: 17.73. Result ID: 698c2bb3112aeb9460efecb0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa644e1e209131cea03e63.md>)

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

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