Getting Hindi text out of a scan

A scanned marksheet, certificate or form is a photograph as far as your computer is concerned. The letters are shapes, not text, which is why you cannot select a word or search for a name. Optical character recognition looks at those shapes and works out which characters they are, and it can do this in Devanagari as well as in the Latin alphabet.

Hindi is harder than English for a recognition engine. Characters join along a continuous headline, matras sit above and below the line, and conjuncts fuse two letters into one shape. A scan that would be perfectly readable in English can come back with mistakes in Hindi, so always read the result rather than trusting it.

Give it the best input you can. Scan at about 300 DPI, keep the page flat and square to the glass, and use even light with no shadow across the text. A photograph taken at an angle costs more accuracy than any setting can recover, so straighten the page before running recognition rather than after.

Pick the tool by what you want back. OCR PDF takes a scanned document and returns the same document with a searchable text layer underneath the picture, so it looks unchanged but can be searched, copied and converted. Image to Text takes a single photograph and hands back the words as plain text. Both offer English, Hindi, or both together for a document that mixes the two.