Traditionally, intelligent document processing relied on OCR and rule-based extraction. These methods could tell ...