I have a problem with the current tesseract I use tesseract in PHP coding
like under
Image Upload Success";
echo '';
shell_exec('"C:\\Program Files (x86)\\Tesseract-OCR\\tesseract.exe"
"D:\\xampp\\htdocs\\Image_OCR\\Image1\\'.$file_name.'" out');
echo "OCR after reading";
$myfile = fo
I have a problem with the current tesseract. I have documents that have
sections of varying background and text colors. Ive read that tesseract v3
was white/black invariant and it didn't matter if I had white text on red
background. But now it matters. The problem is, other parts in the same
im
Yeah my bad seems that just removing the resize code has made it able to
actually detect the 1 as well thanks for all the help.
On Tuesday, May 7, 2019 at 4:04:19 PM UTC-4, zdenop wrote:
>
> Sorry I misread the python message: actually tesseract did not find there
> anything (which at the end i
Sorry I misread the python message: actually tesseract did not find there
anything (which at the end is the same as there is l ;-) )
First of all I would stop with that idiotic resizing. For tesseract is
AFAIR is best to have letter at size 13-30 px. Your as 240!
Try to provide original captured i
I see so should I fool around with the contrast or is there a way to make
it so it while only use number selection when reading the image?
On Tuesday, May 7, 2019 at 3:26:50 PM UTC-4, zdenop wrote:
>
> probably because it is recognized as "l" instead of 1 and you can not
> convert letter to inte
probably because it is recognized as "l" instead of 1 and you can not
convert letter to integer.
Zdenko
ut 7. 5. 2019 o 21:21 Sean Connell
napísal(a):
> Thanks a bunch it works now. The only issue it has is when trying to
> detect the number 1 for some reason it just thinks nothing is there.
>
This is where you need to improve contrast.
https://pillow.readthedocs.io/en/stable/reference/ImageEnhance.html
You need to play a little with PIL to find out what works best for your
data.
Lorenzo
Il giorno mar 7 mag 2019 alle ore 21:21 Sean Connell <
nightfire120sla...@gmail.com> ha scritto:
Thanks a bunch it works now. The only issue it has is when trying to detect
the number 1 for some reason it just thinks nothing is there.
--
You received this message because you are subscribed to the Google Groups
"tesseract-ocr" group.
To unsubscribe from this group and stop receiving emails
you need to add space at the end of tessdata_dir_config because later you
add to it another string with configurations:
tessdata_dir_config = r'--tessdata-dir "S:\Tesseract\tessdata" '
Zdenko
ut 7. 5. 2019 o 19:57 Sean Connell
napísal(a):
> So I added this line of code tessdata_dir_config =
So I added this line of code tessdata_dir_config = r'--tessdata-dir
"S:\Tesseract\tessdata"' and downloaded the English repository from the
link you provided but now I get an error (see attached picture).
On Tuesday, May 7, 2019 at 12:41:33 PM UTC-4, zdenop wrote:
>
> This change was sufficie
This change was sufficient. BTW: I use data from best repository[1]
[1] https://github.com/tesseract-ocr/tessdata_best
Zdenko
ut 7. 5. 2019 o 18:01 Sean Connell
napísal(a):
> Thank you I'll give that a go and see if it works any better. Is it worth
> trying to increase the contrast as well?
Thank you I'll give that a go and see if it works any better. Is it worth
trying to increase the contrast as well?
On Tuesday, May 7, 2019 at 9:52:21 AM UTC-4, zdenop wrote:
>
> modify last part of your code to this:
>
> # invert image and convert to grayscale
> inverted = PIL.ImageOps.invert(new
modify last part of your code to this:
# invert image and convert to grayscale
inverted = PIL.ImageOps.invert(newim2).convert('LA')
loopTest = (pytesseract.image_to_string(
inverted, config=tessdata_dir_config + '--psm 8 --oem 3'))
print(loopTest)
loopTest = int(loopTest)
Do not forget to imp
Thanks a bunch for the response. How would I go about inverting the image and
increasing the contrast though. Sorry I'm still learning how all this works.
--
You received this message because you are subscribed to the Google Groups
"tesseract-ocr" group.
To unsubscribe from this group and stop
I am using tesseract 3.05.00 stable version alongside tesserocr and
couldn't use WordFontAttributes to check whether a word is bold, font-size
etc.. But, I can find the text detection but not size and is bold?
On Friday, May 19, 2017 at 10:39:59 PM UTC+5:30, zdenop wrote:
>
> tesseract 3.05 (the
Hi, try to invert the images (black text on white) and use psm 6 or 7.
Increasing contrast may also help.
Lorenzo
Il mar 7 mag 2019, 08:49 Sean Connell ha
scritto:
> Currently my program searches for the picture of the word Opponents on the
> screen then moves a bit a takes a picture of the
16 matches
Mail list logo
