# tr - Text Recognition
一款针对扫描文档的离线文本识别SDK，核心代码全部采用C++开发，并提供Python接口
#### 更新说明
+ c++接口支持
+ 支持Windows系统
+ 添加python2支持
+ 去除opencv-python、Pillow依赖，降低部署难度
+ 支持多线程
+ 支持GPU
#### Contact
Email: anycode@yahoo.com
#### Requirements
+ Windows 7+/Ubuntu 16.04/Ubuntu 18.04
+ python2/python3，需要安装numpy
+ Windows系统需要安装[VC_redist.x64.exe](https://download.visualstudio.microsoft.com/download/pr/89a3b9df-4a09-492e-8474-8f92c115c51d/B1A32C71A6B7D5978904FB223763263EA5A7EB23B2C44A0D60E90D234AD99178/VC_redist.x64.exe
)
#### GPU版本安装说明
如果对速度有要求，推荐安装GPU版本<br>
要使用GPU版本，复制tr_gpu文件夹里面的文件到tr文件夹<br>
注意: 需要先安装CUDA 10.1以及cuDNN 7.6.5<br>
<br>
若不想安装CUDA/cuDNN，可以使用docker部署
<pre>docker pull mcr.microsoft.com/azureml/onnxruntime:v1.3.0-cuda10.1-cudnn7
sudo nvidia-docker run -v /path/to/tr:/path/to/tr --rm -it mcr.microsoft.com/azureml/onnxruntime:v1.3.0-cuda10.1-cudnn7
</pre>
#### Install
+ 安装方法一
<pre>git clone https://github.com/myhub/tr.git
cd ./tr
sudo python setup.py install
</pre>
+ 安装方法二
<pre>sudo pip install git+https://github.com/myhub/tr.git@master
</pre>
#### Test
<pre>
python2 demo.py # python2兼容测试
python3 test.py # 可视化测试
python3 test-multi-thread.py # 多线程测试
</pre>
#### 关联项目
+ 若需要Web端调用，推荐参考<a href="https://github.com/alisen39/TrWebOCR">TrWebOCR</a>
#### Python Example
<pre>import tr
# detect text lines, return list of (cx, cy, width, height, angle)
print(tr.detect("imgs/web.png", tr.FLAG_RECT))
# detect text lines with angle, return list of (cx, cy, width, height, angle)
print(tr.detect("imgs/id_card.jpeg", tr.FLAG_ROTATED_RECT))
# recognize text line, return (text, confidence)
print(tr.recognize("imgs/line.png"))
# detect and recognize text lines with angle, return list of ((cx, cy, width, height, angle), text, confidence)
print(tr.run("imgs/id_card.jpeg"))
</pre>
#### C++ Example
<pre>tr_init(0, 0, "crnn.bin", NULL);
#define MAX_WIDTH 512
int unicode[MAX_WIDTH];
float prob[MAX_WIDTH];
auto ws = tr_recognize(0, (void *)"line.png", 0, 0, 0, unicode, prob, MAX_WIDTH);
tr_release(0);
</pre>
#### 效果展示
<img src="https://gitee.com/microic/tr/raw/master/imgs/output/id_card/1.png" />
<img src="https://gitee.com/microic/tr/raw/master/imgs/output/id_card/2.png" />
<hr>
<img src="https://gitee.com/microic/tr/raw/master/imgs/output/name_card/1.png" />
<img src="https://gitee.com/microic/tr/raw/master/imgs/output/name_card/2.png" />
tr-master.zip （28个子文件） 
tr-master 
.gitignore 70B README.md 3KB test.py 2KB tr_gpu libonnxruntime.so.1.3.0 87.12MB libtr.so 11.42MB test-multi-thread.py 748B LICENSE 11KB .gitattributes 33B imgs 
web.png 92KB
name_card.jpg 132KB output id_card 1.png 239KB 2.png 20KB name_card 1.png 179KB 2.png 22KB id_card.jpeg 23KB line.png 10KB MANIFEST.in 177B tr crnn.bin 41.86MB ctpn.bin 52.19MB libonnxruntime.so.1.3.0 10.08MB libtr.dll 36.34MB onnxruntime.dll 5.8MB __init__.py 34B tr.py 6KB libtr.so 11.42MB
char_table.txt 20KB setup.py 868B demo.py 295B
评论0