This project is besed on text-detection-ctpn and sceneReco.Because I don't have a Nvidia gpu,so this project is run on cpu mode.
-
ubuntu16.04
-
anaconda3
-
opencv3
-
tensorflow
-
pytorch
-
warp_ctc_pytorch
CTPN:Detecting Text in Natural Image with Connectionist Text Proposal Network
There has three directories and a file at the root directory of this project:
type | name | description |
---|---|---|
directory | ctpn | the module of ctpn |
directory | crnn | the module of crnn |
directory | imgs | the images that will be processed |
file | demo.py | the main function of this project |
Download the model for crnn from baidu yun:here.Then put it into crnn/models
Run install-cpu.sh , if there have some problem when run it , please open the file and execute single sentense.
There are two methods to run this project.
-
If you have many images to be recognized,you can put them into imgs/ and run python demo.py
-
If you only have one or two images to be recognized,you can just run python demo.py image1 image2 ...