# Scene Text Recognition on Standard STR Benchmark

Source: [Scene Text Recognition on Standard STR Benchmark](<https://www.sota2.com/research/sota/scene-text-recognition-on-standard-str-benchmark>)

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

Data updated: 2026-02-26T05:52:26.893Z

Evaluation of Scene Text Recognition performance on the Standard STR Benchmark, measured using word accuracies based on character length (36, 62, and 94 characters).

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 | Scene Text Recognition |
| Dataset (including any stated split) | Standard STR Benchmark |
| Metric | Mean Word Accuracy (36-char) |
| Unit | PERCENT |
| 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

- [Mean Word Accuracy (36-char)](<https://www.sota2.com/research/sota/scene-text-recognition-on-standard-str-benchmark.md?metric=Mean%20Word%20Accuracy%20(36-char)>): Higher is better; unit: PERCENT.

- [Mean Word Accuracy (62-char)](<https://www.sota2.com/research/sota/scene-text-recognition-on-standard-str-benchmark.md?metric=Mean%20Word%20Accuracy%20(62-char)>): Higher is better; unit: PERCENT.

- [Mean Word Accuracy (94-char)](<https://www.sota2.com/research/sota/scene-text-recognition-on-standard-str-benchmark.md?metric=Mean%20Word%20Accuracy%20(94-char)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Mean Word Accuracy (36-char) | Result date | Report paper |
| --- | --- | --- | --- |
| PARSeq\_A | 96 | Not recorded | 68fa6d031e209131cea37551 |
| TRBA | 95.2 | Not recorded | 68fa6d031e209131cea37551 |
| ABINet | 95.2 | Not recorded | 68fa6d031e209131cea37551 |
| PARSeq\_N | 95.2 | Not recorded | 68fa6d031e209131cea37551 |
| ViTSTR-S | 94.3 | Not recorded | 68fa6d031e209131cea37551 |
| PARSeq\_A | 91.9 | Not recorded | 68fa6d031e209131cea37551 |
| PARSeq\_N | 90.7 | Not recorded | 68fa6d031e209131cea37551 |
| TRBA | 90.6 | Not recorded | 68fa6d031e209131cea37551 |
| ABINet | 89.8 | Not recorded | 68fa6d031e209131cea37551 |
| ViTSTR-S | 88.6 | Not recorded | 68fa6d031e209131cea37551 |
| CRNN | 88.5 | Not recorded | 68fa6d031e209131cea37551 |
| CRNN | 83.2 | Not recorded | 68fa6d031e209131cea37551 |

### PARSeq\_A

Mean Word Accuracy (36-char): 96. Result ID: 698f78b03a6c314849f8ddae.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Implementation](<https://github.com/baudm/parseq>)

Recorded evaluation configuration: {"Train data":"R"}. This record does not assert independent reproduction.

### TRBA

Mean Word Accuracy (36-char): 95.2. Result ID: 698f78b03a6c314849f8ddab.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa626b1e209131ce9f94bb.md>)

Recorded evaluation configuration: {"Train data":"R"}. This record does not assert independent reproduction.

### ABINet

Mean Word Accuracy (36-char): 95.2. Result ID: 698f78b03a6c314849f8ddac.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa697f1e209131cea227bf.md>)

[Implementation](<https://github.com/FangShancheng/ABINet>)

Recorded evaluation configuration: {"Train data":"R"}. This record does not assert independent reproduction.

### PARSeq\_N

Mean Word Accuracy (36-char): 95.2. Result ID: 698f78b03a6c314849f8ddad.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Implementation](<https://github.com/baudm/parseq>)

Recorded evaluation configuration: {"Train data":"R"}. This record does not assert independent reproduction.

### ViTSTR-S

Mean Word Accuracy (36-char): 94.3. Result ID: 698f78b03a6c314849f8ddaa.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69ef1e209131cea250ab.md>)

[Implementation](<https://github.com/roatienza/deep-text-recognition-benchmark>)

Recorded evaluation configuration: {"Train data":"R"}. This record does not assert independent reproduction.

### PARSeq\_A

Mean Word Accuracy (36-char): 91.9. Result ID: 698f78b03a6c314849f8dda8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Implementation](<https://github.com/baudm/parseq>)

Recorded evaluation configuration: {"Train data":"S"}. This record does not assert independent reproduction.

### PARSeq\_N

Mean Word Accuracy (36-char): 90.7. Result ID: 698f78b03a6c314849f8dda7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Implementation](<https://github.com/baudm/parseq>)

Recorded evaluation configuration: {"Train data":"S"}. This record does not assert independent reproduction.

### TRBA

Mean Word Accuracy (36-char): 90.6. Result ID: 698f78b03a6c314849f8dda5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa626b1e209131ce9f94bb.md>)

Recorded evaluation configuration: {"Train data":"S"}. This record does not assert independent reproduction.

### ABINet

Mean Word Accuracy (36-char): 89.8. Result ID: 698f78b03a6c314849f8dda6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa697f1e209131cea227bf.md>)

[Implementation](<https://github.com/FangShancheng/ABINet>)

Recorded evaluation configuration: {"Train data":"S"}. This record does not assert independent reproduction.

### ViTSTR-S

Mean Word Accuracy (36-char): 88.6. Result ID: 698f78b03a6c314849f8dda4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69ef1e209131cea250ab.md>)

[Implementation](<https://github.com/roatienza/deep-text-recognition-benchmark>)

Recorded evaluation configuration: {"Train data":"S"}. This record does not assert independent reproduction.

### CRNN

Mean Word Accuracy (36-char): 88.5. Result ID: 698f78b03a6c314849f8dda9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a211e209131cea26339.md>)

Recorded evaluation configuration: {"Train data":"R"}. This record does not assert independent reproduction.

### CRNN

Mean Word Accuracy (36-char): 83.2. Result ID: 698f78b03a6c314849f8dda3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d031e209131cea37551.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a211e209131cea26339.md>)

Recorded evaluation configuration: {"Train data":"S"}. This record does not assert independent reproduction.

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