Re: [Bug-ocrad] The function ignore_wide_blobs() doth ignore too much, methinks

[Antonio Diaz Diaz]

Tilman Hausherr wrote:
> Then I tested with production. I generally got better results (many
> pages now do have useful output that didn't before), with one exception.
> One of the images was a huge b/w photograph, thus, from an OCR point of
> view, a huge amount of noise. The OCR needed several minutes (!). Thus,
> although I can't look into your mind, I guess that the "if" statement
> was probably meant as a safety against exactly that. However, that
> safety measure also prevents the OCR of printed excel tables with grey
> background cells. So instead of making a an assumption about individual
> areas, OCRAD ignores the whole file.

Ocrad can waste a lot of time trying to create text lines from noise 
just to produce a lot of garbage. This is what the "if" statement tries 
to prevent.

But in this case the problem is in image preprocessing (before feeding 
it to ocrad). Those "grey background cells" are not grey at all, but 
full of speckles. This works well enough for the human eye, but makes 
life difficult to OCR programs.

Just see what one of those "grey background cells" looks like to ocrad:

.......O.......O..O....O..O.O..O..O....O..O....O..O....O..O....O..O.O..O....
...O..O....O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O...
..O....O..O....O..OO...O..O.O..O..O....O..O.O..O..O....O..O.O..O..O.O..O..O.
...O..O.O..O..O.O..O..O....O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O...
.......O.......O..O....O..O.O..O.......O..O....O..O.O..O..O.O..O..O.O..O....
.O.O..O..O.O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O...
.......OOOOOOOOOOOOOOO.O..O.O..O..O.O..O..O.O..O..O.O..O.OOOO..O..O.O..O..O.
O..O..OOOOOOOOOOOOOOOOO.O..O..O.O..O..O.O..O..O.O..OOOO.OOOOOOOOOOOOO.O.O..O
....O..OOOOOOOOOOOOOOO.O.......O..O....O..O.O..O..OOO..O.OOOOOOOOOOOO..O..O.
O..OO.O....O.OOOO..O..O.O..O..O.OO.O..O.O..O..O.O.OOO.O.OOOOOOOOOOOOO.O.O..O
..O.O..O.....OOO....O..O....O..O.......O..O....O..OOO..O....O..O.OOOO..O..O.
OO.O..O.OO.O.OOOO..O..O.O..OOOOOO..O.OOOO..O..OOOOOOOOO.O..O..O.OOOO..O.O...
..O....O.....OOO..O....O.OOOOOOOOOO.OOOOO.O..OOOOOOOOOOO..O....OOOO....O..O.
O..O....O..O.OOOO..O..O.OOOOOOOOOOOO..OOOO.OOOOOOOOOOOO.O..O..OOOO.O..O.O..O
....O..O....OOOO..O.O..OOOOOO..OOOO...OOOOO.OOOO.OOOOOOO..O....OOO.....O..O.
...O..O....O.OOOO..O..OOOO.O..O.OOOO..OOOOOOOOO.O.OOO.O.O..O..OOO..O..O.O...
.......O.....OOO....O..OOOOOOOOOOOOO...OOOOOOO.O..OOO..O..O...OOO.O.O..O..O.
O..O..O..O.O.OOOO..O..OOOOOOOOOOOOOO..O.OOOOO.O.O..OO.O.O..O.OOOO..O..O.O..O
O.O....O.....OOO......OOOOOOOOOOOOOO...O.OOOO..O..OOO..O....OOOO..O....O..O.
O..O..O....O.OOOO..O..OOOO.O....O..O..O.OOOOOOO.O.OOO.O.O..OOOOOO..O..O.O..O
..O....O..O..OOO..O....OOOO....O.OO.O..OOOOOOO.O..OOO..O..O.OOOO..O.O..O..O.
O..O..O.O..O.OOOO..O..OOOOOO..O.OOOO..OOOO.OOOO.O.OOO.O.O..OOOO.O..O..O.O..O
.......O.....OOO....O..OOOOOOOOOOOO.O.OOOO..OOOO..OOOO.O..O.OOOO..O.O..O....
...O..O....O.OOOO..O..O.OOOOOOOOOOOO.OOOO..OOOOOO.OOOOO.O..OOOO.O..O..O.O..O
....O..O..O..OOO..O.O..O.OOOOOOOOOO.OOOO..O.OOOOO.OOOOOO..OOOOOO..O....O..O.
...O..O.O..O..O.O..O..O.O..OOOO.O..O..O.O..O..O.O..OOOO.O..OOOO.O..O..O.O...
....O..O.......O..O.O..O..O....O.......O..O.O..O..O.O..O..O.O..O..O.O..O..O.
.O.O..O.OO.O..O.OO.O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O...
..O....O..O....O..O....O..O....O..O.O..O..O.O..O..O.O..O..O.O..O....O..O..O.
...O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O..O.O...
.......O..O....O..O....O..O.O..O..O.O..O..O.O..O..O.O..O..O.O..O....O..O..O.

Processing images for human consumption is very different from 
processing them for OCR. The right way to handle non-white backgrounds 
in the latter case is binarizing the image using a suitable threshold or 
feeding the image to ocrad as a greymap or colormap.

For images already processed as the one above an option telling ocrad 
not to remove noise inside a wide blob may help in some cases.


Regards,
Antonio.