rezazad68 / bcdunet_dibco Goto Github PK

Documnet Image Binarization, DIBCO Challenges

Python 100.00%

computer-vision deep-learning deep-neural-networks document-image-binarization handwriting handwriting-recognition historical-documents image-binarization image-segmentation keras-tensorflow machine-learning pre-processing

bcdunet_dibco's Issues

How can i test your model "Learned weights" on my image ?

Hi
how can i run your pretrained model on my data ?
thank you

Please provide your packages' version

I am doing something related to OCR and Document Binarization is very important. I tried your code, but it's failed to run demo. Please provide your packages' version to re-produce your experiments. Thank you so much!

How could I evaluate single image

Have you tried UNet?

The input and output image sizes are not identical

Hello, Reza,
Thank you for your wonderful BCDUnet_DIBCO project. I found that the sizes of the input and output images are not identical. Maybe that bug lies in the extract_ordered_overlap or recompone_overlap. Please fix it, and thanks again~

how to save the original image and output image? Such as the images folder

Hi, excuse me,I want to know how to save the original image and output image? Such as the images folder? but I run the evaluate.py that find that does not output the image.So thanks for your answer!

Recommend Projects