# Handwriting Recognition on IAM Paragraph Level (test)

Source: [Handwriting Recognition on IAM Paragraph Level (test)](<https://www.sota2.com/research/sota/handwriting-recognition-on-iam-paragraph-level-test>)

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

Data updated: 2026-02-26T05:52:25.347Z

Evaluation of Handwriting Recognition performance on the test split of the IAM Paragraph Level dataset, measured by Character Error Rate (CER) both with and without Language Model constraints.

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 | Handwriting Recognition |
| Dataset (including any stated split) | IAM Paragraph Level (test) |
| Metric | Mean Test CER (w/o LM) |
| Unit | PERCENT |
| Direction | Lower 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

- [Mean Test CER (w/o LM)](<https://www.sota2.com/research/sota/handwriting-recognition-on-iam-paragraph-level-test.md?metric=Mean%20Test%20CER%20(w%2Fo%20LM)>): Lower is better; unit: PERCENT.

- [Corpus CER (w/o LM)](<https://www.sota2.com/research/sota/handwriting-recognition-on-iam-paragraph-level-test.md?metric=Corpus%20CER%20(w%2Fo%20LM)>): Lower is better; unit: PERCENT.

- [Mean Test CER (w/ LM)](<https://www.sota2.com/research/sota/handwriting-recognition-on-iam-paragraph-level-test.md?metric=Mean%20Test%20CER%20(w%2F%20LM)>): Lower is better; unit: PERCENT.

- [Corpus CER (w/ LM)](<https://www.sota2.com/research/sota/handwriting-recognition-on-iam-paragraph-level-test.md?metric=Corpus%20CER%20(w%2F%20LM)>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Mean Test CER (w/o LM) | Result date | Report paper |
| --- | --- | --- | --- |
| FPHR+Aug | 6.3 | Not recorded | 68fa6cd71e209131cea36299 |
| FPHR | 6.7 | Not recorded | 68fa6cd71e209131cea36299 |
| SLS + MDLSTM + CTC | 7.5 | Not recorded | 68fa6cd71e209131cea36299 |
| Bluche | 7.9 | Not recorded | 68fa6cd71e209131cea36299 |
| Bluche | 10.1 | Not recorded | 68fa6cd71e209131cea36299 |
| SLS + MDLSTM + CTC | 11.1 | Not recorded | 68fa6cd71e209131cea36299 |
| Bluche et al. | 16.2 | Not recorded | 68fa6cd71e209131cea36299 |

### FPHR+Aug

Mean Test CER (w/o LM): 6.3. Result ID: 698f42783a6c314849f784ff.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cd71e209131cea36299.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cd71e209131cea36299.md>)

Recorded evaluation configuration: {"Resolution":"~145 dpi","Data Augmentation":true}. This record does not assert independent reproduction.

### FPHR

Mean Test CER (w/o LM): 6.7. Result ID: 698f42783a6c314849f784fe.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cd71e209131cea36299.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cd71e209131cea36299.md>)

Recorded evaluation configuration: {"Resolution":"~145 dpi"}. This record does not assert independent reproduction.

### SLS + MDLSTM + CTC

Mean Test CER (w/o LM): 7.5. Result ID: 698f42783a6c314849f78503.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cd71e209131cea36299.md>)

Recorded evaluation configuration: {"Resolution":"300 dpi","Paragraph Segmentation Required":true,"Line Segmentation":"SLS"}. This record does not assert independent reproduction.

### Bluche

Mean Test CER (w/o LM): 7.9. Result ID: 698f42783a6c314849f78501.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cd71e209131cea36299.md>)

Recorded evaluation configuration: {"Resolution":"300 dpi","Paragraph Segmentation Required":true}. This record does not assert independent reproduction.

### Bluche

Mean Test CER (w/o LM): 10.1. Result ID: 698f42783a6c314849f78500.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cd71e209131cea36299.md>)

Recorded evaluation configuration: {"Resolution":"150 dpi","Paragraph Segmentation Required":true}. This record does not assert independent reproduction.

### SLS + MDLSTM + CTC

Mean Test CER (w/o LM): 11.1. Result ID: 698f42783a6c314849f78502.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cd71e209131cea36299.md>)

Recorded evaluation configuration: {"Resolution":"150 dpi","Paragraph Segmentation Required":true,"Line Segmentation":"SLS"}. This record does not assert independent reproduction.

### Bluche et al.

Mean Test CER (w/o LM): 16.2. Result ID: 698f42783a6c314849f784fd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6cd71e209131cea36299.md>)

Recorded evaluation configuration: {"Resolution":"150 dpi"}. This record does not assert independent reproduction.

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