Handwritten Text Recognition on Belfort
Metric: WER (%) (lower is better)
Metric: WER (%) (lower is better)
| # | Model↕ | WER (%)▲ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | PyLaia (all transcriptions + agreement-based split) | 15.14 | No | Handwritten Text Recognition from Crowdsourced A... | 2023-06-19 | - |
| 2 | PyLaia (rover consensus + agreement-based split) | 17.08 | No | Handwritten Text Recognition from Crowdsourced A... | 2023-06-19 | - |
| 3 | PyLaia (human transcriptions + agreement-based split) | 19.12 | No | Handwritten Text Recognition from Crowdsourced A... | 2023-06-19 | - |
| 4 | PyLaia (human transcriptions + random split) | 28.11 | No | Handwritten Text Recognition from Crowdsourced A... | 2023-06-19 | - |