# Face Verification on Oulu-CASIA

Source: [Face Verification on Oulu-CASIA](<https://www.sota2.com/research/sota/face-verification-on-oulu-casia>)

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

Data updated: 2026-02-26T05:27:33.620Z

Evaluation of Face Verification performance on the Oulu-CASIA dataset, reporting accuracy.

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) | Oulu-CASIA |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 10 of 10 |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/face-verification-on-oulu-casia.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Proposed dynamic MTL | 99.14 | Not recorded | 68fa66991e209131cea113d4 |
| Static MTL | 98 | Not recorded | 68fa66991e209131cea113d4 |
| Single-task (fine-tuning) | 97.71 | Not recorded | 68fa66991e209131cea113d4 |
| FaceNet | 97.5 | Not recorded | 68fa66991e209131cea113d4 |
| DeepID-2,3 | 96.5 | Not recorded | 68fa66991e209131cea113d4 |
| SphereFace | 95.5 | Not recorded | 68fa66991e209131cea113d4 |
| Naive dynamic MTL | 95.14 | Not recorded | 68fa66991e209131cea113d4 |
| Centerloss | 95.1 | Not recorded | 68fa66991e209131cea113d4 |
| VGGFace | 93.5 | Not recorded | 68fa66991e209131cea113d4 |
| Single-task (pretrained) | 92.6 | Not recorded | 68fa66991e209131cea113d4 |

### Proposed dynamic MTL

Accuracy: 99.14. Result ID: 698cfb038b517eb4331e24a0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66991e209131cea113d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66991e209131cea113d4.md>)

[Implementation](<https://github.com/hengxyz/Dynamic_multi-task-learning.git>)

Recorded evaluation configuration: {"Images":"1.1M"}. This record does not assert independent reproduction.

### Static MTL

Accuracy: 98. Result ID: 698cfb038b517eb4331e249e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66991e209131cea113d4.md>)

Recorded evaluation configuration: {"Images":"1.1M"}. This record does not assert independent reproduction.

### Single-task (fine-tuning)

Accuracy: 97.71. Result ID: 698cfb038b517eb4331e249d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66991e209131cea113d4.md>)

Recorded evaluation configuration: {"Images":"1.1M"}. This record does not assert independent reproduction.

### FaceNet

Accuracy: 97.5. Result ID: 698cfb038b517eb4331e2498.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66991e209131cea113d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62ed1e209131ce9fbe20.md>)

Recorded evaluation configuration: {"Images":"200M"}. This record does not assert independent reproduction.

### DeepID-2,3

Accuracy: 96.5. Result ID: 698cfb038b517eb4331e2497.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66991e209131cea113d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa619c1e209131ce9f515e.md>)

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

### SphereFace

Accuracy: 95.5. Result ID: 698cfb038b517eb4331e249b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66991e209131cea113d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63f41e209131cea01c92.md>)

[Implementation](<https://github.com/wy1iu/sphereface>)

Recorded evaluation configuration: {"Images":"0.7M"}. This record does not assert independent reproduction.

### Naive dynamic MTL

Accuracy: 95.14. Result ID: 698cfb038b517eb4331e249f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66991e209131cea113d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64071e209131cea02185.md>)

Recorded evaluation configuration: {"Images":"1.1M"}. This record does not assert independent reproduction.

### Centerloss

Accuracy: 95.1. Result ID: 698cfb038b517eb4331e249a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66991e209131cea113d4.md>)

Recorded evaluation configuration: {"Images":"0.7M"}. This record does not assert independent reproduction.

### VGGFace

Accuracy: 93.5. Result ID: 698cfb038b517eb4331e2499.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66991e209131cea113d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa61d01e209131ce9f62f5.md>)

[Implementation](<http://www.robots.ox.ac.uk/~vgg/research/very_deep/>)

Recorded evaluation configuration: {"Images":"2.6M"}. This record does not assert independent reproduction.

### Single-task (pretrained)

Accuracy: 92.6. Result ID: 698cfb038b517eb4331e249c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa66991e209131cea113d4.md>)

Recorded evaluation configuration: {"Images":"1.1M"}. This record does not assert independent reproduction.

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