Mar-12-2023, 07:11 AM
Hi,
I am refining , = making more efficient, my programs for
OCR-ing Prayer Cards. I do not know what PCs look like in your countries,
but imagine a small, foulded document = it has 4 printable sides.
Only one of those sides has the content we need to OCR. (all sides are scanned)
So I crop the document into 4 pieces, and examine them one by one,
until I find where the money is. That works, but I could optimize.
Questions: (a "page" is 1/4 of the PC)
1) Sometimes a page is blank, nothing on it.
2) Sometimes a page is completely covered with an image(not half..., completely)
OCR spends time trying to find specific words on those pages, when it finds none, it goes to the next .
How could I determine that a page is blank, or fully covered with colors or grey values ,
so I can skip it right away? (Preferably not by examining the RGB value of each pixel)
Any ideas?,
Paul
I am refining , = making more efficient, my programs for
OCR-ing Prayer Cards. I do not know what PCs look like in your countries,
but imagine a small, foulded document = it has 4 printable sides.
Only one of those sides has the content we need to OCR. (all sides are scanned)
So I crop the document into 4 pieces, and examine them one by one,
until I find where the money is. That works, but I could optimize.
Questions: (a "page" is 1/4 of the PC)
1) Sometimes a page is blank, nothing on it.
2) Sometimes a page is completely covered with an image(not half..., completely)
OCR spends time trying to find specific words on those pages, when it finds none, it goes to the next .
How could I determine that a page is blank, or fully covered with colors or grey values ,
so I can skip it right away? (Preferably not by examining the RGB value of each pixel)
Any ideas?,
Paul