|
--- |
|
language: |
|
- fa |
|
library_name: hezar |
|
tags: |
|
- image-to-text |
|
- hezar |
|
pipeline_tag: image-to-text |
|
--- |
|
|
|
A CRNN model for Persian OCR. This model is based on a simple CNN + LSTM architecture inspired by [this paper](https://arxiv.org/abs/1507.05717). |
|
More info about data and training will be provided soon. |
|
|
|
Note that this model is only optimized for scanned documents and supports input characters of up to 32 (For an end-to-end OCR pipeline, use a text detector |
|
model first to extract text boxes preferrably in word-level and then use this model), but it can be used to be fine-tuned on other domains like license plate |
|
or handwritten texts. |
|
|
|
#### Limitations |
|
This model is best suited for Persian alphabet and lacks the ability to recognize numbers and digits properly. We'll soon retrain this model to fit all scenarios. |
|
|
|
#### Usage |
|
``` |
|
pip install hezar |
|
``` |
|
|
|
```python |
|
from hezar.models import Model |
|
|
|
crnn = Model.load("hezarai/crnn-base-fa-64x256") |
|
texts = crnn.predict(["sample_image.jpg"]) |
|
print(texts) |
|
``` |
