Cnn-dedge: multilingual scene text detection and extraction

HIGHLIGHTS

• What: The authors propose an algorithm customized architecture with bit plane for identifying the textual regions from natural image samples. The authors have tested the models on MSRA TD500, ICDAR 2017, MRRC and MLe2e data sets and results are compared with other deep learning models Test snake, EAST, segment link, pixel link, PAN, baseline CNN and ECN etc. The model is implemented with a CRNN approach to identify the textual data in scene samples and significantly decreased the quantity of factors compared to prior scene textual data identification models. Unlike the approaches discussed above, which use filter . . .
  

   

  

  If you want to have access to all the content you need to log in!

  Thanks :)

  Username or Email

  Password

  Remember me

  Lost your password?

  If you don't have an account, you can create one here.