OCR When Searchable Does Not Mean Accurate

OCR & TEXTQUALITY · AGUIDEFORUKLAWYERS OCR: When Searchable Does Not Mean Accurate Scanned Documents, Recognition Errors and the Text Layer Your Searches Actually Run Against COMPUTERFORENSICSLAB § ABOUTTHEAUTHOR PRE PA REDBYCOMPUTERFORENSICSLABE - DISCOVERYTEAM SEARCH VALIDATION CPR PART 35 EXPERT REPORTS FULL CHAIN-OF-CUSTODY DOCUMENTATION § CONTENTS In this guide 01 Executive summary 02 The problem in plain English: pictures of words 03 How OCR works and how it fails 04 Measuring text quality: scoring, triage and re-OCR 05 The searching consequences: terms against imperfect text 06 The hard cases: handwriting, faxes, stamps and legacy paper 07 OCR in the protocol: postures, disclosure and disputes 08 Source architecture: where else the evidence lives 09 Worked examples 10 Common mistakes and technical limitations 11 Questions to ask · Suggested wording 12 Checklist and red flags · When to involve a digital forensic expert 13 Frequently asked questions 14 Glossary · References · Disclaimer · How a specialist laboratory can assist § 0 1 · ORIENTATION Executive summary THEHEADLINEPOINT : OCRTEXTISANESTIMATE : MEASUREIT , TRIAGEIT , ANDNEVER CERTIFYASEARCHOVERTEXTNOBODYSCORED § 0 2 · FIRSTPRINCIPLES The problem in plain English: pictures of words § 0 3 · THEMACHINERY How OCR works and how it fails § 0 4 · THEMEASUREMENT Measuring text quality: scoring, triage and re-OCR § 0 5 · THECONSEQUENCES The searching consequences: terms against imperfect text § 0 6 · THEHARDCASES The hard cases: handwriting, faxes, stamps and legacy paper § 0 7 · THEPAPERWORK OCR in the protocol: postures, disclosure and disputes § 0 8 · THEWIDERMAP Source architecture: where else the evidence lives WHEN THE SCAN COUNTERPART EARLIER STRUCTURED HUMAN DEFEATS COPIES GENERATIONS RECORDS READING OCR BORN- RENDERED / DIGITAL QUOTED ORIGINAL FORMS § 0 9 · INTHEWILD Worked examples EXAMPLE1 · THEGUARANTEEFOUNDBYITSOWNMISREADING EXAMPLE2 · THECERTIFICATEOVERSTATIC EXAMPLE3 · THEHANDWRITTENMARGINTHATOUTWEIGHEDTHETYPEDPAGE § 1 0 · WHEREITGOESWRONG Common mistakes and technical limitations Common mistakes Technical limitations § 1 1 · INTERROGATORIES & DRAFTINGAIDS Questions to ask · Suggested wording Ask your client Ask your opponent Ask your eDiscovery / forensic provider SUGGESTEDWORDING · OCRPOSTURE PA R AG RAPHFORTHEDRD / PROTOCOL § 1 2 · QUICKCONTROL Checklist and red flags · When to involve a digital forensic expert The OCR checklist Red flags When to involve a digital forensic expert § 1 3 · COMMONQUESTIONS Frequently asked questions Our platform says the documents are searchable. Isn't that enough? How accurate is modern OCR really? Searches found nothing in our scanned archive. Can we rely on that? Can OCR read handwriting, signatures and stamps? Should we re-OCR an archive that was processed years ago? The other side produced scans whose text garbles the key clause. What do we do? § 1 4 · REFERENCE Glossary Sources and authoritative references DISCLAIMER § HOWASPECIALISTLABORATORYCANASSIST Working with Computer Forensics Lab Speak to a forensic examiner, not a salesperson. INSTRUCTTHELAB NEWENQUIRIESEMAILE - DISCOVERY