You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
67 lines
1.6 KiB
Markdown
67 lines
1.6 KiB
Markdown
# ocrd_calamari
|
|
|
|
> Recognize text using [Calamari OCR](https://github.com/Calamari-OCR/calamari).
|
|
|
|
[](https://circleci.com/gh/OCR-D/ocrd_calamari)
|
|
[](https://pypi.org/project/ocrd_calamari/)
|
|
[](https://codecov.io/gh/OCR-D/ocrd_calamari)
|
|
|
|
## Introduction
|
|
|
|
This offers a OCR-D compliant workspace processor for some of the functionality of Calamari OCR.
|
|
|
|
This processor only operates on the text line level and so needs a line segmentation (and by extension a binarized
|
|
image) as its input.
|
|
|
|
## Installation
|
|
|
|
### From PyPI
|
|
|
|
```
|
|
pip install ocrd_calamari
|
|
```
|
|
|
|
### From Repo
|
|
|
|
```sh
|
|
pip install .
|
|
```
|
|
|
|
## Install models
|
|
|
|
Download standard models:
|
|
|
|
```
|
|
wget https://github.com/Calamari-OCR/calamari_models/archive/master.zip
|
|
unzip master.zip
|
|
```
|
|
|
|
Download models trained on GT4HistOCR data:
|
|
|
|
```
|
|
wget https://file.spk-berlin.de:8443/calamari-models/GT4HistOCR/model.tar.xz
|
|
mkdir gt4hist-calamari
|
|
cd gt4hist-calamari
|
|
tar xf ../model.tar.xz
|
|
```
|
|
|
|
## Example Usage
|
|
|
|
~~~
|
|
ocrd-calamari-recognize -p test-parameters.json -m mets.xml -I OCR-D-SEG-LINE -O OCR-D-OCR-CALAMARI
|
|
~~~
|
|
|
|
With `test-parameters.json`:
|
|
~~~
|
|
{
|
|
"checkpoint": "/path/to/some/trained/models/*.ckpt.json"
|
|
}
|
|
~~~
|
|
|
|
TODO
|
|
----
|
|
|
|
* Support Calamari's "extended prediction data" output
|
|
* Currently, the processor only supports a prediction using confidence voting of multiple models. While this is
|
|
superior, it makes sense to support single model prediction, too.
|