FOTS: Fast Oriented Text Spotting with a Unified Network

Introduction

Do text detect and OCR in an entire End-to-End model

All detect and OCR modules have been tested!

Installation

• tensorflow
• opencv

Train

python2 multigpu_train.py [gpu_list]

You should give the path to the dataset labeled in ICDAR format in file 'FOTS/dataset/dataReader.py'

Test

call FOTS_testModel()

Author

Call me yufeng_yu@sina.com

From @SF.Tech