Update README.md
Browse files
README.md
CHANGED
@@ -25,39 +25,43 @@ This model performs Handwritten Text Recognition in Norwegian. It was developed
|
|
25 |
|
26 |
## Model description
|
27 |
|
28 |
-
The model has been trained using the PyLaia library on the [NorHand v1](https://zenodo.org/record/6542056)
|
29 |
|
30 |
Training images were resized with a fixed height of 128 pixels, keeping the original aspect ratio.
|
31 |
|
32 |
-
|
|
33 |
-
|
|
34 |
| train | 19,653 |
|
35 |
| val | 2,286 |
|
36 |
| test | 1,793 |
|
37 |
|
38 |
-
An external 6-gram character language model can be used to improve recognition. The language model is trained on the text from the NorHand v1 training set.
|
39 |
-
|
40 |
## Evaluation results
|
41 |
|
42 |
The model achieves the following results:
|
43 |
|
44 |
-
| set | Language model | CER (%) | WER (%) |
|
45 |
|:------|:---------------| ----------:| -------:|----------:|
|
46 |
| test | no | 7.94 | 24.04 | 1,793 |
|
47 |
| test | yes | 6.55 | 18.20 | 1,793 |
|
48 |
|
|
|
|
|
49 |
## How to use?
|
50 |
|
51 |
-
Please refer to the [documentation](https://atr.pages.teklia.com/pylaia/).
|
52 |
|
53 |
# Cite us!
|
54 |
|
55 |
```bibtex
|
56 |
-
@inproceedings{
|
57 |
-
author =
|
58 |
-
title =
|
59 |
-
booktitle =
|
60 |
-
year =
|
|
|
|
|
|
|
|
|
61 |
}
|
62 |
```
|
63 |
|
|
|
25 |
|
26 |
## Model description
|
27 |
|
28 |
+
The model has been trained using the PyLaia library on the [NorHand v1](https://zenodo.org/record/6542056) dataset.
|
29 |
|
30 |
Training images were resized with a fixed height of 128 pixels, keeping the original aspect ratio.
|
31 |
|
32 |
+
| set | horizontal lines |
|
33 |
+
| :---- | ------: |
|
34 |
| train | 19,653 |
|
35 |
| val | 2,286 |
|
36 |
| test | 1,793 |
|
37 |
|
|
|
|
|
38 |
## Evaluation results
|
39 |
|
40 |
The model achieves the following results:
|
41 |
|
42 |
+
| set | Language model | CER (%) | WER (%) | lines |
|
43 |
|:------|:---------------| ----------:| -------:|----------:|
|
44 |
| test | no | 7.94 | 24.04 | 1,793 |
|
45 |
| test | yes | 6.55 | 18.20 | 1,793 |
|
46 |
|
47 |
+
An external 6-gram character language model can be used to improve recognition. The language model is trained on the text from the NorHand v1 training set.
|
48 |
+
|
49 |
## How to use?
|
50 |
|
51 |
+
Please refer to the [PyLaia documentation](https://atr.pages.teklia.com/pylaia/usage/prediction/) to use this model.
|
52 |
|
53 |
# Cite us!
|
54 |
|
55 |
```bibtex
|
56 |
+
@inproceedings{pylaia2024,
|
57 |
+
author = {Tarride, Solène and Schneider, Yoann and Generali-Lince, Marie and Boillet, Mélodie and Abadie, Bastien and Kermorvant, Christopher},
|
58 |
+
title = {{Improving Automatic Text Recognition with Language Models in the PyLaia Open-Source Library}},
|
59 |
+
booktitle = {Document Analysis and Recognition - ICDAR 2024},
|
60 |
+
year = {2024},
|
61 |
+
publisher = {Springer Nature Switzerland},
|
62 |
+
address = {Cham},
|
63 |
+
pages = {387--404},
|
64 |
+
isbn = {978-3-031-70549-6}
|
65 |
}
|
66 |
```
|
67 |
|