arturaugusto/display_ocr

Real-time image preprocess and OCR.

Display OCR

Unmaintained. Alternative traineddata avaliable at https://github.com/Shreeshrii/tessdata_ssd or alternative implementation that don't use tesseract: https://github.com/arturaugusto/7seg-ocr.

---

OpenCV-Python + python-tesseract real-time image preprocess and OCR.

Trained data for 7 segments font avaliable under letsgodigital folder.

Web-app using trained data: http://ocr.sytes.net (Not always up, due to the low cost server)

Reference: https://code.google.com/p/python-tesseract/

Tips to achieve better results:

• Use erode to avoid gaps between the segments.
• Avoid direct light on the display (I use parchment paper to diffuse the light)

License: GPL v2

display_ocr's Issues

Training

Hi,

I have compiled Tesseract 3.05.2 with Leptonica 1.76.0.

When I try to run your trainig:
python2 tesseract-trainer.py

I get very small output letsgodigital.tesstrain file, 144 kb.

And errors like these:

read_params_file: Can't open nobatch
read_params_file: Can't open box.train.stderr
Tesseract Open Source OCR Engine v3.05.02 with Leptonica
Empty page!!
Empty page!!

P.S. I set properly TESSDATA_PREFIX to your folder ./letsgodigital/

Question is s Tesseract 3.02 necessary for the training?

getting "Please call SetImage before attempting recognition" warning/error

I'm using OS X Mavericks with:
tesseract 3.03
leptonica-1.70
libjpeg 8d : libpng 1.6.12 : zlib 1.2.5

opencv: stable 2.4.9

How to run it?

I have installed the files via git clone. I don't know how to use it.
Please do help, thanks

Installation issue

Can you help me how to use this?

0 is always recognized as 8

I'm using the traineddata file from this project to recognize seven segment digits. Even for a very clear & high-resolution sample, 0 is recognized as 8. I searched through some stackoverflow, issue seems to be common. Take a look.
http://stackoverflow.com/questions/30479002/digital-numbers-on-tesseract-ocr

Update to python 3?

Hey,
Is possible for this to be updated to work in python 3? I've had a play with the getting it to work but it seems that the wrapper for tesseracts API is broken.

Thanks.

problem with OCR.py

Hi ! I have a problem with OCR.py (ubuntu) :

lorisson@lorisson-MS-7677:/usr/lib/python2.7$ /home/lorisson/Bureau/display_ocr-master/OCR.py
Traceback (most recent call last):
  File "/home/lorisson/Bureau/display_ocr-master/OCR.py", line 104, in <module>
    thresh = ConfigSectionMap("PREPROCESS")['threshold']
  File "/home/lorisson/Bureau/display_ocr-master/OCR.py", line 17, in ConfigSectionMap
    options = Config.options(section)
  File "/usr/lib/python2.7/ConfigParser.py", line 279, in options
    raise NoSectionError(section)
ConfigParser.NoSectionError: No section: 'PREPROCESS'

Recognizing 26 but not 43 for me

Sorry, this isn't really an issue, more of a plea for help. I am executing the follow statement from python on the two attached images:

text = pytesseract.image_to_string(Image.open(filename), lang="letsgodigital", boxes=False, config="digits")

The "26" works fine but the "43" doesn't come back with a result. Anything I could try? I tried dilating the image further but no luck. Is it because the "43" is slightly rotated? Perhaps because the 3 is too close to the edge of the image?

How to install and run it?

@arturaugusto Hi, I want to know the project's runtime environment.

some error in tesseract or OCR.py?

I install python-opencv and python-teressact and run OCR.py. This is my result:

nano@nano-MOV:$ cd display_ocr-master/
nano@nano-MOV:/display_ocr-master$ python OCR.py
VIDIOC_QUERYMENU: Invalid argument
VIDIOC_QUERYMENU: Invalid argument
VIDIOC_QUERYMENU: Invalid argument
Error opening data file ./tessdata/letsgodigital.traineddata
Please make sure the TESSDATA_PREFIX environment variable is set to the parent directory of your "tessdata" directory.
Failed loading language 'letsgodigital'
Tesseract couldn't load any languages!

the program is going on but when i want to select a region the application stop:

Traceback (most recent call last):
File "OCR.py", line 203, in
Recognize(iplimage)
File "OCR.py", line 39, in Recognize
full_text = api.GetUTF8Text()
File "/usr/lib/python2.7/dist-packages/tesseract.py", line 10556, in
getattr = lambda self, name: _swig_getattr(self, TessBaseAPI, name)
File "/usr/lib/python2.7/dist-packages/tesseract.py", line 57, in _swig_getattr
raise AttributeError(name)
AttributeError: GetUTF8Text