Count the number of characters in an image
6 views (last 30 days)
Show older comments
I want to count the number of letters present in the image..How to code it? I just want to count,not recognizing by ocr... Please help me with the code. Thank you
0 Comments
Answers (2)
Walter Roberson
on 25 Nov 2015
If you do not use OCR then you cannot determine where the boundaries are between letters. For example on the 6th line near the end, the spacing is enough between the left stroke of the 'H' and the right-stroke together with horizontal bar, that without ocr one would tend to count that as two different letters. Then by using average width one might perhaps break apart the horizontal bar and right vertical stroke of the 'H' from the following 'e', but if you do so then you would have arrived at 3 characters, the left stroke of the H, the horizontal and right stroke of the H, and the 'e'. It requires OCR to recognize that the horizontal and right stroke does not form a character and that you need to "back up" to the left stroke to piece together an 'H'.
Therefore the task cannot be done accurately without using OCR.
3 Comments
Image Analyst
on 25 Nov 2015
See the section on handwriting analysis on this page: http://www.visionbib.com/bibliography/contentschar.html#OCR,%20Document%20Analysis%20and%20Character%20Recognition%20Systems
I think Walter's idea of identifying connected components and then binning the width to determine if the blob is 1, 2, 3 or whatever characters long is worth trying. I was going to suggest a similar thing based on the area. Either way, you'll have a problem in that some letters, like an i, are narrower and have less area than others, say a w.
0 Comments
See Also
Categories
Find more on Language Support in Help Center and File Exchange
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!