How to check text you got from an image
Reading text from an image is a big time saver, but you should always check the result. Sinhala and Tamil have many letters that look alike, so small errors slip in.
Where mistakes usually happen
- Similar letters. Letters with small differences in a curve or a dot get swapped.
- Vowel signs. The little marks above, below or beside a letter are easy to drop or change.
- Numbers and dates. Always check these against the original.
- Names. Names can't be guessed from context, so they are the most likely to be wrong.
- Mixed language. Lines that mix English with Sinhala or Tamil sometimes get split up oddly.
A simple way to check
Put the image and the result side by side and read them line by line. Do it on the whole text, not just the parts that look odd.
Getting a better result next time
A sharper image gives fewer mistakes. See getting text out of an image for tips on photos and scans.