[python3.6] 运用tf实现自然场景文字检测,keras/pytorch实现ctpn+crnn+ctc实现不定长场景文字OCR识别

A distributed visual search and visual data analytics platform.

Perform image segmentation and background removal in javascript using superpixes

Implementation of our paper 'PixelLink: Detecting Scene Text via Instance Segmentation' in AAAI2018

Convolutional Recurrent Neural Network (CRNN) for image-based sequence recognition.

Extremely simple implement for CRNN by Tensorflow

text detection mainly based on ctpn model in tensorflow, id card detect, connectionist text proposal network

CRNN with attention to do OCR,add Chinese recognition

yolo3+ocr

基于CTPN（tensorflow）+CRNN（pytorch）+CTC的不定长文本检测和识别

Chinese text detection and recognition based on CTPN + DENSENET using Keras and Tensor Flow，使用keras和tensorflow基于CTPN+Densenet实现的中文文本检测和识别

CNN+LSTM+CTC based OCR implemented using tensorflow.

use ctpn detect id cards and printed transfer orders then crnn discern them

CTPN + DenseNet + CTC based end-to-end Chinese OCR implemented using tensorflow and keras

用keras实现OCR定位、识别

ctpn+crnn Scene character recognition

本项目实现了ocr主流算法gru/lstm+ctc+cnn架构，进行不定长度验证码识别，达到不分割字符而识别验证码内容的效果。验证码内容包含了大小字母以及数字，并增加点、线、颜色、位置、字体等干扰项。本项目对gru +ctc+cnn、lstm+ctc+cnn、cnn三种架构进行了对比，实践说明同等训练下gru/lstm+ctc+cnn架构准确率和速度均明显优于cnn架构，gru +ctc+cnn优于lstm+ctc+cnn，在实验2500个样本数据200轮训练时，gru +ctc+cnn架构在500样本测试准确率达90.2%。本项目技术能够训练长序列的ocr识别，更换数据集和相关调整，即可用于比如身份证号码、车牌、手机号、邮编等识别任务，也可用于汉字识别。

A tensorflow implementation of EAST text detector

ocr, cnn+lstm+ctc, crnn, recognition model, tensorflow

AdvancedEAST is an algorithm used for Scene image text detect, which is primarily based on EAST, and the significant improvement was also made, which make long text predictions more accurate.