OCR (text recognition)

Technology extracting text from images or PDF. PDF → DOCX often uses OCR. Quality depends on original clarity and language.