Generally finding text on color images is very hard, but in your case, your business card image is in binary, so I have two possible solutions in my mind:
1. You can use OpenCV function findContours() which will detect contours of letters, then you can filter that contours (using hierarchy). After that you put bounding rect on that filtered contours. By now you have coordinates of letters and shapes on you business card. Based on that information you can group the bounding rects to contain your blocks of texts. 2. On your binary image there are only white and black colors. So you have 2 cases, where letters and shapes are white and background black or where letters and shapes are black and background white. You can use histogram to detect the color of letters. Next you can use histogram on every row of image and on every column of image and see where are the blocks of text. On Monday, February 3, 2014 11:41:30 PM UTC+1, Nick Porter wrote: > > Thanks for the reply Aleksander. These improved the accuracy of my scans, > however it does not provide a sloution to detecting paragraphs and blocks > of text. Any idea how to do this? > > On Monday, February 3, 2014 2:16:15 AM UTC-7, Aleksander Grzyb wrote: >> >> To improve results you should try to: >> >> 1. Convert image to binary image. >> 2. Crop the image to get rid off the surroundings. >> 3. Detect skew of image and do some perspective transform. >> >> I recommend to use OpenCV to do this operations. There is a pod for >> OpenCV: >> >> https://github.com/Fl0p/OpenCV-iOS >> >> Here are some links that should help you do the image processing part: >> >> >> http://stackoverflow.com/questions/8667818/opencv-c-obj-c-detecting-a-sheet-of-paper-square-detection >> >> http://stackoverflow.com/questions/6555629/algorithm-to-detect-corners-of-paper-sheet-in-photo >> >> http://stackoverflow.com/questions/8637867/skew-detection-and-reduction-in-opencv >> >> http://stackoverflow.com/questions/7838487/executing-cvwarpperspective-for-a-fake-deskewing-on-a-set-of-cvpoint >> >> W dniu piątek, 31 stycznia 2014 20:44:55 UTC+1 użytkownik Nick Porter >> napisał: >>> >>> I am trying to scan a business card using tesseract OCR, all I am doing >>> is sending the image in with no per-prossesing, heres the code I am using. >>> >>> Tesseract* tesseract = [[Tesseract alloc] initWithLanguage:@"eng+ita"]; >>> tesseract.delegate = self; [tesseract >>> setVariableValue:@"0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ@.-()" >>> forKey:@"tessedit_char_whitelist"]; [tesseract setImage:[UIImage >>> imageNamed:@"card.jpg"]]; //image to check [tesseract recognize]; >>> NSLog(@"Here is the text %@", [tesseract recognizedText]); >>> >>> Picture of card <http://imgur.com/nQPG6iq> >>> >>> This is the output <http://imgur.com/poikzBn> >>> >>> As you can see the accuracy is not 100%, which is not what I am >>> concerned about I figure I can fix that with some simple per-processing. >>> However if you notice it mixes the two text blocks at the bottom, which >>> splits up the address, and possibly other information on other cards. >>> >>> How can I possibly use Leptonica(or something else) to group the text >>> somehow? Possibly send regions of text on the image individually to >>> tesseract to scan? I've been stuck on this problem for a while any possible >>> solutions are welcome! >>> >> -- -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/tesseract-ocr?hl=en --- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
