Implement Vision API OCR for computer vision tasks. Extract image text with `TEXT_DETECTION` or `DOCUMENT_TEXT_DETECTION` for dense documents and handwriting.