EAST: An Efficient and Accurate Scene Text Detector

Introduction

This is a PyTorch re-implementation of EAST: An Efficient and Accurate Scene Text Detector (paper).

The features are summarized blow:

• Only RBOX part is implemented.
• A fast Locality-Aware NMS in C++ provided by the paper's author.
• The pre-trained model provided achieves 81.61 F1-score on ICDAR 2015 Incidental Scene Text Detection Challenge using only training images from ICDAR 2015 and 2013. see here for the detailed results.
• Differences from original paper
  • Use ResNet-50 rather than PVANET
  • Use dice loss (optimize IoU of segmentation) rather than balanced cross entropy
  • Use linear learning rate decay rather than staged learning rate decay

Performance

ICDAR 2015

Model	Recall	Precision	Hmean	Download
PyTorch re-implementation of EAST	74.48%	90.26%	81.61%	Link

Link

Offline evaluation

$ python eval.py
$ ./eval.sh

Credit

Most codes are ported from argman/EAST (the Tensorflow re-implementation).

DataSet

Model is trained & tested on ICDAR 2015. Please download following 4 files then put them under "data" folder:

• ch4_training_images.zip
• ch4_training_localization_transcription_gt.zip
• ch4_test_images.zip
• Challenge4_Test_Task1_GT.zip

Dependency

• PyTorch 1.1.0

Usage

Data Pre-processing

Extract training & test images:

$ python extract.py

Train

$ python train.py

If you want to visualize during training, run in your terminal:

$ tensorboard --logdir runs

Demo

Pick 10 random test examples from ICDAR-2015:

$ python demo.py

Examples