Ufdjrw

Is there a good way to store credentials outside of a password manager?

Valid Badminton Score?

What are the ramifications of creating a homebrew world without an Astral Plane?

What defines a dissertation?

Where in the Bible does the greeting ("Dominus Vobiscum") used at Mass come from?

Trouble understanding overseas colleagues

Coordinate position not precise

There is only s̶i̶x̶t̶y one place he can be

Why are on-board computers allowed to change controls without notifying the pilots?

Is it correct to write "is not focus on"?

Time travel short story where a man arrives in the late 19th century in a time machine and then sends the machine back into the past

Is a roofing delivery truck likely to crack my driveway slab?

What to do with wrong results in talks?

Print name if parameter passed to function

Tiptoe or tiphoof? Adjusting words to better fit fantasy races

Why does John Bercow say “unlock” after reading out the results of a vote?

Modify casing of marked letters

How will losing mobility of one hand affect my career as a programmer?

Have I saved too much for retirement so far?

How does residential electricity work?

Why Were Madagascar and New Zealand Discovered So Late?

Transcription Beats per minute

How does a character multiclassing into warlock get a focus?

Irreducibility of a simple polynomial

Remove background of image containing text

Finding the index of an item given a list containing it in PythonDoes Python have a string 'contains' substring method?image processing to improve tesseract OCR accuracyImage Processing: Algorithm Improvement for 'Coca-Cola Can' RecognitionHow do I choose between Tesseract and OpenCV?OCR: Image to text?How to extracting line of text from image using OpenCVImprove quality of tesseract text recognitionImprove Tesseract OCR results with blurred textWhat's the way to remove all lines and borders in image(keep texts) programmatically?

2

1

I am building custom ocr for some documents. After getting ROI I am passing them to tesseract. To improve accuracy I want to remove background of image. I am observing that when there are images like this:

tesseract is not able to read anything.(Because of lines in the image)

But for images like this: Its giving correct results.
Can any one suggest how to remove everything from image except text?

python opencv image-processing ocr tesseract

share|improve this question

edited Mar 11 at 6:50

008karan

asked Mar 7 at 11:35

008karan008karan

284

• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  An heuristic approach would be to keep only the color of the text (and possibly a small region around them) and make the remaining white for example. Not sure it will work though.
  
  – Eypros
  Mar 7 at 11:43
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  That might work. Any guide how to do that?
  
  – 008karan
  Mar 7 at 13:11
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  @Eypros Loved the nerd calculator on your profile. You should start with thresholding the image to see if you can get rid of the colors.
  
  – Rick M.
  Mar 7 at 13:42
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  I tried thresolding but I am getting B&W image with those lines in it
  
  – 008karan
  Mar 7 at 14:04
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  With the free ocr api I get "JTFF" for the image, out of the box. Often using an Asian language like korean works better for tricky English(!) letters. So you can try this.
  
  – Jim Grigoryan
  Mar 7 at 19:23
  
  

  
  
  
  

add a comment | 

2

1

I am building custom ocr for some documents. After getting ROI I am passing them to tesseract. To improve accuracy I want to remove background of image. I am observing that when there are images like this:

tesseract is not able to read anything.(Because of lines in the image)

But for images like this: Its giving correct results.
Can any one suggest how to remove everything from image except text?

python opencv image-processing ocr tesseract

share|improve this question

edited Mar 11 at 6:50

008karan

asked Mar 7 at 11:35

008karan008karan

284

• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  An heuristic approach would be to keep only the color of the text (and possibly a small region around them) and make the remaining white for example. Not sure it will work though.
  
  – Eypros
  Mar 7 at 11:43
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  That might work. Any guide how to do that?
  
  – 008karan
  Mar 7 at 13:11
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  @Eypros Loved the nerd calculator on your profile. You should start with thresholding the image to see if you can get rid of the colors.
  
  – Rick M.
  Mar 7 at 13:42
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  I tried thresolding but I am getting B&W image with those lines in it
  
  – 008karan
  Mar 7 at 14:04
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  With the free ocr api I get "JTFF" for the image, out of the box. Often using an Asian language like korean works better for tricky English(!) letters. So you can try this.
  
  – Jim Grigoryan
  Mar 7 at 19:23
  
  

  
  
  
  

add a comment | 

I am building custom ocr for some documents. After getting ROI I am passing them to tesseract. To improve accuracy I want to remove background of image. I am observing that when there are images like this:

tesseract is not able to read anything.(Because of lines in the image)

But for images like this: Its giving correct results.
Can any one suggest how to remove everything from image except text?

python opencv image-processing ocr tesseract

share|improve this question

edited Mar 11 at 6:50

008karan

asked Mar 7 at 11:35

008karan008karan

284

share|improve this question

edited Mar 11 at 6:50

008karan

asked Mar 7 at 11:35

008karan008karan

284

share|improve this question

edited Mar 11 at 6:50

008karan

asked Mar 7 at 11:35

008karan008karan

284

asked Mar 7 at 11:35

008karan008karan

284

asked Mar 7 at 11:35

008karan008karan

284