这里的EAST便是旷视在CVPR2017的一篇文章：EAST: An Efficient and Accurate Scene Text Detector的实现，

不过用训练好的模型文件frozen_east_text_detection.pb进行文本检测的结果还是不够好，总有一些文字未能检测出来

Rough Google Translate translation of the above text (26-Feb-2019 - JRE:

The EAST here is an escapist article in CVPR2017: The implementation of EAST: An Efficient and Accurate Scene Text Detector,

However, the result of text detection using the trained model file frozen_east_text_detection.pb is still not good enough, and some texts are not detected.

Referenced article: https://arxiv.org/abs/1704.03155v2