someoneAlready / east-text-detection-with-mxnet

EAST: An Efficient and Accurate Scene Text Detector

EAST: An Efficient and Accurate Scene Text Detector

This is a mxnet version of An Efficient and Accurate Scene Text Detector project

Introduction

This is a mxnet re-implementation of EAST: An Efficient and Accurate Scene Text Detector. The features are summarized blow:

• Only RBOX part is implemented.
• A fast Locality-Aware NMS in C++ provided by the paper's author.
• 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
• Speed on 720p (resolution of 1280x720) images:
  • Now
    • Graphic card: TiTan X
    • Network fprop: ~34 ms
    • NMS (C++): ~6ms
    • Overall: ~24 fps

Please cite his paper if you find this useful.

Other Implementations

Thank you all for the efforts! If you have new implementation and want to share with others, feel free to make a pull request or email me!

1. the origin github implementations tensorflow

Contents

1. Installation
2. Download
3. Test
4. Train
5. Examples

Installation

1. Any version of mxnet version > 0.12.0 should be ok.

Download

1. Pre-trained Model on ICDAR 2015 (training set): BaiduYun link
2. ICDAR 2015 dataset: BaiduYun link

Train

If you want to train the model, you should provide the dataset path, in the dataset path, a separate gt text file should be provided for each image and run train_mx.py file

If you have more than one gpu, you can set gpu ids to gpu_list(like'0,1,2,3')

Test

run

python2 demo.py

a text file will be then written to the output path.

Examples

Here are some test examples on icdar2015, enjoy the beautiful text boxes!

Please let me know if you encounter any issues(my email [e-mail] and my wechat wp2572201844).