OCR not reading characters properly from this image

Question

I am trying to read characters from the captcha image using OCR engine. Unfortunately, I am unable to get a proper result. The image is being read as

:
18<?.< rest of the content is trimmed due to evaluation restriction>.

Which OCR I should implement and use in Java so that every character is read with 100% accuracy.

Captcha are usually meant not to be read by OCR, so I'm not surprised you got something wrong.. — user2340612, May 28 '18 at 08:29
It is hard for an OCR engine to read from CAPTCHA. Still, refer: https://stackoverflow.com/questions/1813881/java-ocr-implementation — Gaurav, May 28 '18 at 08:29
AFIK, there is no OCR program with **100%** accuracy...(So far) — zlakad, May 28 '18 at 08:30

score 1 · Answer 1 · answered May 28 '18 at 08:58

Looking at the image, it looks like a captcha. Instead of reading it as a whole you can segment the letters first using a library like OpenCV (link).(Maybe the findContours() method) Once you have every letter seperately, you can apply your OCR and then combine the results obtained.

OCR not reading characters properly from this image

1 Answers1