Detect punctuation marks on ocr

Hi,
I'm trying to do an application regarding ocr.
How can the ocr detect punctuation marks like period?
i trained an image in the "ocr trainer" the ocr trainer detected the numbers and period, but "period" is not recognized when i tried to write the output to .txt file
here is the sample processed image
i already processed the original image using the following code snippet
str = "testimage.jpg";
I = imread('testRawImage.jpg');
I = rgb2gray(I);
Icorrected = imtophat(I,strel('cube',8));
contrastAdjusted = imadjust(Icorrected);
BW1 = imbinarize(contrastAdjusted);
imwrite(BW1,str)
output in txt file is (period is not recognized, but i checked on ocr trainer the period is recognized using the ocr trainer):
1420545
1420575
142024
142031
i used fprintf to write in txt file.
if it helps i used the language i made using the ocr trainer.
this is my ocr snippet
ocrResults = ocr(H, 'Language','test.traineddata');

4 comentarios

Image Analyst
Image Analyst el 26 de Abr. de 2020
What are H and 'test.traineddata'? Please attach the script so we can try it ourselves. And attach the original image rather than the processed image.
Robert Cadavos
Robert Cadavos el 26 de Abr. de 2020
hi i uploaded the files i used for the application, unfortunately I cannot upload the .txt file and traineddate file because of the file format.
test.traineddata is the output of ocr trainer
Robert Cadavos
Robert Cadavos el 26 de Abr. de 2020
i got it working i just changed the ocr to
ocrResults = ocr(H,'CharacterSet','.0123456789');
Thanks

Iniciar sesión para comentar.

Respuestas (0)

Categorías

Más información sobre MATLAB Mobile en Centro de ayuda y File Exchange.

Etiquetas

Preguntada:

el 26 de Abr. de 2020

Comentada:

el 26 de Abr. de 2020

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by