Hi Lorenzo,
Thanks for detailed description of pre-processing steps. I will link from
the wiki so that it is available for easy reference.
Thank you for sharing.
--
You received this message because you are subscribed to the Google Groups
"tesseract-ocr" group.
To unsubscribe from this group and
Hi,
yes, at the very least you can use some adaptive threshold method, like
OTSU, to find the best parameters. But OTSU has its own parameters so you
need to fine tune those too (a little).
What worked best for me was first to do a rough normalization of the images
(lightness, contrast) and then d
Hi Shree,
I'd love to but it is a commercial project I'm working on so I cannot share
the current solution.
I will try to find the old scripts I used for the first attempts. Basically
it was something like this:
- normalize lightness
- make illumination uniform (CLAHE on HSV "V" channel)
- denois
Thank you so much for your sharing.
It seems a very complicated cleanup. It will be very useful if you can
provide some preprocessing script. And I am wondering there is also some
thresholds depending on different images, right?
By the way, I have read some papers about LSTM +Ctc for ocr. The
ad
Hi Lorenzo,
Do you have a script for image pre-processing? Please share, if possible.
It will be helpful to many.
On Wed, Apr 3, 2019 at 6:47 PM Lorenzo Bolzani wrote:
> Hi, I train with real data. I use grayscale images, I think color makes no
> difference.
>
> I do a very good image cleanup:
Hi, I train with real data. I use grayscale images, I think color makes no
difference.
I do a very good image cleanup: background removal, denoise, straightening,
sharpening, illumination correction, contrast stretching, etc. before
passing the text to tesseract. This part is likely better done o
Thank you for your kind reminder. I have done this. And confusing thing
happens. I train my model with grey scale image without any 300dpi resize.
It goes not that well when I validate with my some of my test data. But if
I resize test image to 300 dpi, it’s better. In fact ,image will be resize
to
I haven't trained with real images. I would guess that training images
should be similar to what you will be using for OCR. It might be best to
test with a small set of images and see what works best for you.
On Wed, Apr 3, 2019 at 2:38 PM Du Kotomi wrote:
> If we use text2image tool, there is n
If we use text2image tool, there is no such problem.
What about training with our real data. I have enough images for training.
Should I need to do some preprocess like binary or resized dpi and then do
lstm training?
On Wed, Apr 3, 2019 at 16:36 Shree Devi Kumar wrote:
> Usually for LSTM train
Usually for LSTM training we are using synthetic images created by
text2image program using training text and fonts using tesstrain.sh or
tesstrain.py. Hence there is no question of binarization or dpi as the
program creates images as expected by tesseract training process.
On Wed, Apr 3, 2019 at
Anybody here?
On Wed, Apr 3, 2019 at 09:57 wrote:
> Sorry for disturb again. I have sent my issue befire, but no one gives the
> answer. I really need your help.
>
>
> I go through the source code and find tesseract do Otsu Thresholding and
> put the binary pix in the Thresholder object.
> But
Sorry for disturb again. I have sent my issue befire, but no one gives the
answer. I really need your help.
I go through the source code and find tesseract do Otsu Thresholding and
put the binary pix in the Thresholder object.
But It seems the Thresholder object haven't been invoked if I us
I go through the source code and find tesseract do Otsu Thresholding and
put the binary pix in the Thresholder object.
But It seems the Thresholder object haven't been invoked if I use lstm
engines.
As well as dpi size,tesseract wiki said it is better for 300 dpi. This is a
requirement for tes
13 matches
Mail list logo
