# Latin typeface generation (cross-reconstruction) on Proprietary Latin typeface dataset 52 letters (test)

Source: [Latin typeface generation (cross-reconstruction) on Proprietary Latin typeface dataset 52 letters (test)](<https://www.sota2.com/research/sota/latin-typeface-generation-cross-reconstruction-on-proprietary-latin-typeface-dataset-52-letters-test>)

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

Data updated: 2026-04-09T23:28:20.007Z

Evaluation of Latin typeface generation models using cross-reconstruction techniques on the proprietary 52-letter dataset (test split), assessed using IoU, L1, RE, Cont. Acc., and Align. Acc.

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 | Latin typeface generation (cross-reconstruction) |
| Dataset (including any stated split) | Proprietary Latin typeface dataset 52 letters (test) |
| Metric | mIoU |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [mIoU](<https://www.sota2.com/research/sota/latin-typeface-generation-cross-reconstruction-on-proprietary-latin-typeface-dataset-52-letters-test.md?metric=mIoU>): Higher is better; unit: PERCENT.

- [L1 Loss](<https://www.sota2.com/research/sota/latin-typeface-generation-cross-reconstruction-on-proprietary-latin-typeface-dataset-52-letters-test.md?metric=L1%20Loss>): Lower is better; unit: SCALAR.

- [RE](<https://www.sota2.com/research/sota/latin-typeface-generation-cross-reconstruction-on-proprietary-latin-typeface-dataset-52-letters-test.md?metric=RE>): Lower is better; unit: SCALAR.

- [Continuous Accuracy](<https://www.sota2.com/research/sota/latin-typeface-generation-cross-reconstruction-on-proprietary-latin-typeface-dataset-52-letters-test.md?metric=Continuous%20Accuracy>): Higher is better; unit: PERCENT.

- [Alignment Accuracy](<https://www.sota2.com/research/sota/latin-typeface-generation-cross-reconstruction-on-proprietary-latin-typeface-dataset-52-letters-test.md?metric=Alignment%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | mIoU | Result date | Report paper |
| --- | --- | --- | --- |
| DeepVecFont-v2 | 73.5 | Not recorded | 69d732787d0172140ec76a82 |
| DesigNet | 71.1 | Not recorded | 69d732787d0172140ec76a82 |
| DesigNet | 69.3 | Not recorded | 69d732787d0172140ec76a82 |
| DeepVecFont-v2 | 68.1 | Not recorded | 69d732787d0172140ec76a82 |
| DeepVecFont-v2 | 67.5 | Not recorded | 69d732787d0172140ec76a82 |
| DualVector | 56.4 | Not recorded | 69d732787d0172140ec76a82 |

### DeepVecFont-v2

mIoU: 73.5. Result ID: 69d8360347f34ce924bff345.

[Reporting paper](<https://www.sota2.com/research/paper/69d732787d0172140ec76a82.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6eea1e209131cea42dec.md>)

Recorded evaluation configuration: {"Number of shots":"10"}. This record does not assert independent reproduction.

### DesigNet

mIoU: 71.1. Result ID: 69d8360347f34ce924bff343.

[Reporting paper](<https://www.sota2.com/research/paper/69d732787d0172140ec76a82.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d732787d0172140ec76a82.md>)

Recorded evaluation configuration: {"Number of shots":"1","Self-Refinement module usage":"false"}. This record does not assert independent reproduction.

### DesigNet

mIoU: 69.3. Result ID: 69d8360347f34ce924bff344.

[Reporting paper](<https://www.sota2.com/research/paper/69d732787d0172140ec76a82.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d732787d0172140ec76a82.md>)

Recorded evaluation configuration: {"Number of shots":"1","Self-Refinement module usage":"true"}. This record does not assert independent reproduction.

### DeepVecFont-v2

mIoU: 68.1. Result ID: 69d8360347f34ce924bff341.

[Reporting paper](<https://www.sota2.com/research/paper/69d732787d0172140ec76a82.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6eea1e209131cea42dec.md>)

Recorded evaluation configuration: {"Number of shots":"1"}. This record does not assert independent reproduction.

### DeepVecFont-v2

mIoU: 67.5. Result ID: 69d8360347f34ce924bff342.

[Reporting paper](<https://www.sota2.com/research/paper/69d732787d0172140ec76a82.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6eea1e209131cea42dec.md>)

Recorded evaluation configuration: {"Number of shots":"1","Self-Refinement module usage":"true"}. This record does not assert independent reproduction.

### DualVector

mIoU: 56.4. Result ID: 69d8360347f34ce924bff340.

[Reporting paper](<https://www.sota2.com/research/paper/69d732787d0172140ec76a82.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f521e209131cea4508d.md>)

[Implementation](<https://github.com/thuliu-yt16/dualvector>)

Recorded evaluation configuration: {"Number of shots":"1"}. This record does not assert independent reproduction.

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