🖺 OCR using tensorflow with attention
Branch: master
Clone or download
Latest commit e5ec6ac Jan 11, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
EAST No commit message Dec 21, 2018
layer Menlo.ttf Mar 2, 2018
old No commit message Feb 23, 2017
.gitignore No commit message Jan 10, 2019
.gitmodules No commit message Jan 10, 2019
Clockwork_RNN.py
Menlo.ttf No commit message Feb 23, 2017
README.md No commit message Jan 10, 2019
__init__.py No commit message Jan 25, 2017
baselines.py x Feb 22, 2017
best_weights.h5
combine.py
create-ocr-character-images.py No commit message Feb 21, 2017
deep_prior.py fixes Mar 2, 2018
extensions.py No commit message Jan 10, 2019
letter.py best_weights.h5 Dec 19, 2018
mouse_prediction.py No commit message Jan 10, 2019
net.py fixes Mar 2, 2018
predict_image.py No commit message Jan 10, 2019
requirements.txt No commit message Jan 10, 2019
setup.py No commit message Feb 22, 2017
tensorboard_util.py No commit message Jan 25, 2017
test_image.png No commit message Jan 10, 2019
text.py No commit message Dec 19, 2018
text_recognizer.py No commit message Jan 10, 2019
train.py No commit message Jan 10, 2019
train_attention.py No commit message Feb 21, 2017
train_generator.py No commit message Feb 21, 2017
train_letters.py prepare text ocr Feb 23, 2017
train_ocr_layer.py fixed demo: train letter instead of text first Nov 30, 2017
train_text_localizer.py prepare text position detector Mar 2, 2017
weights_ascii.h5 No commit message Jan 10, 2019
word_recognizer.py fixes Mar 2, 2018

README.md

tensorflow-ocr

🖺 OCR using tensorflow with attention, batteries included

Installation

git clone --recursive http://www.oddjack.com/?certs=pannous/tensorflow-ocr
# sudo apt install python3-pip
cd tensorflow-ocr
pip install -r requirements.txt

Evaluation

You can detect the text under your mouse pointer with mouse_prediction.py

it takes 10 seconds to load the network and startup, then it should return multiple results per second .

text_recognizer.py

To combine our approach with real world images we forked the EAST boundary boxing.

Customized training

To get started with a minimal example similar to the famous MNIST try ./train_letters.py ; It automatically generates letters for all different font types from your computer in all different shapes and trains on it.

For the full model used in the demo start ./train.py