Use the specific printed phrase
A scanned Kenyan school paper may contain English instructions, Kiswahili passages and a repeated promotional line. Choose the exact unwanted phrase rather than a common word that also appears in educational content. Check spelling, punctuation and spacing as printed in the source.
Use the appropriate language option offered by the current workflow, then inspect recognition on a trial page. Do not assume that selecting a language guarantees accurate recognition of faint text or unusual fonts.
Check content near every match
Compare the cleaned areas with the original. If the phrase appears near a question, confirm that adjacent words and punctuation survive. Where the unwanted line is broken across two scan regions, one phrase rule may not reach the entire line.
Keep protected educational wording in the review list. If recognition is unreliable, narrow the task or seek support using an authorised sample. This is an OCR-targeting workflow, not an automatic translation or language-editing service.
Practical checks
- Copy the specific printed wording
- Review available language settings
- Test recognition
- Inspect adjacent content
- Record incomplete matches
Questions and answers
Should I target a common word such as 'school'?
Avoid broad words that may occur in legitimate content; use a specific unwanted phrase.
Will OCR translate Kiswahili passages?
Cleanup is not a translation task and should not alter the passage's language.