ID Analyzer
← ブログに戻る
Document VerificationKYC Policy

Handling Expired, Damaged and Worn ID Documents: When to Accept, Flag or Request a Rescan

ID Analyzer TeamNov 16, 2026読了 5 分
Handling Expired, Damaged and Worn ID Documents: When to Accept, Flag or Request a Rescan

Real-world identity documents are not pristine. Driver licenses crack in wallets, passport laminates peel, and many customers don't notice their ID has expired until onboarding tells them. Your verification flow will see these documents every day, so you need a clear policy for each case before it reaches production.

This post is about that policy layer. It covers what to do once the system has read a document and found it expired, physically damaged or worn, and how to decide between accepting it, flagging it for review or asking for a new capture.

Three Problems That Look Similar but Are Not

Teams often treat expired, damaged and worn documents as one "bad document" bucket. They need different handling.

  • Expired documents may be in perfect condition and fully authentic. The issue is regulatory or contractual validity.
  • Damaged documents have physical defects such as cracks, holes, missing corners, delamination or water damage. These can hide security features or look like tampering.
  • Worn documents have gradual degradation such as faded printing, scuffed holograms or a scratched MRZ. These mainly affect how reliably the data can be read and checked.

Each one fails for a different reason, so each one should trigger a different action.

Expired Documents: A Policy Question, Not a Technical One

Reading the expiry date is the easy part. OCR pulls it from the visual zone, and on many documents it can be cross-checked against the MRZ or a PDF417 barcode. Whether to accept the document is a business and compliance decision.

Questions to answer in your policy

  • Does your regulatory framework allow expired IDs at all? Some onboarding contexts require a currently valid document. Others allow a recently expired one for identity confirmation alone.
  • Is there a grace period? If you allow one, define it explicitly in days and apply it the same way across document types.
  • Does the purpose matter? Confirming age for a low-risk purchase is different from opening a regulated financial account.
  • Are there jurisdiction-specific rules? Some issuing authorities extend validity periods during processing backlogs. If you plan to honor these, document the source and review it regularly.

Run expiry as a deterministic rule after extraction, separate from authenticity checks. A document can be authentic and expired, or valid and forged. Keeping the two signals apart makes your audit trail clearer and stops one result from hiding the other.

If you reject on expiry, tell the user directly: "This document expired on [date]. Please use a current ID." Don't send them into a rescan loop. Rescanning an expired document will never fix it.

Damaged Documents: Separate Damage From Tampering

Damage is where risk and conversion pull hardest against each other. A cracked laminate over the photo might be an old card, or it might be a sign that someone lifted the laminate to swap the photo.

Signals worth weighing

  • Where the damage is. Damage across the portrait, date fields or MRZ deserves more scrutiny than a bent corner on a blank area.
  • Data consistency. If the visual zone, MRZ and barcode still agree, the damage probably hasn't changed the data. If they disagree, treat the document as suspicious no matter how innocent the damage looks.
  • Authentication results. Document authentication and anti-forgery checks assess whether expected features and layout are present for the detected template. A damaged document may fail some checks for physical reasons, so review the specific failures instead of relying on a single pass/fail result.
  • Biometric match. A strong face match between the document portrait and a live selfie with liveness detection gives useful supporting evidence. A weak match on a damaged portrait should go to manual review.

Heads up

Never auto-accept a document just because the damage looks accidental. If damage covers a security-critical area and data or authentication checks are inconsistent, route it to manual review or ask for a different document.

Worn Documents: Usually a Capture Problem First

Worn documents are often perfectly valid but harder to read. Faded ink, scuffed holograms and scratched MRZ lines can lower extraction confidence or cause authentication checks to return inconclusive results.

The first response should usually be a guided rescan, not a rejection. Specific guidance works best:

  • Ask the user to move to diffuse light, since glare on worn laminates makes faded fields harder to read.
  • Ask for the back of the document if it has a barcode, which gives you a second data source when the front is degraded.
  • Limit rescans to two or three attempts. After that, offer an alternative document or manual review.

If repeated captures still return low-confidence fields, the document itself may be too worn to verify automatically. Escalate it. Don't keep asking.

A Simple Decision Matrix

Here is a starting point you can adapt to your risk appetite:

ConditionData consistent?AuthenticationAction
Expired, beyond policyAnyAnyReject with clear reason
Expired, within grace periodYesPassAccept, log exception
Minor damage, non-critical areaYesPassAccept
Damage over photo, data or MRZYesPartialManual review
Any damageNoAnyFlag as suspicious
Worn, low read confidenceUnclearInconclusiveGuided rescan, then review

Make the Policy Auditable

Whatever thresholds you choose, record the reason for each decision along with the result. When an auditor asks why a recently expired license was accepted, "within documented 30-day grace policy, authentication passed, face match passed" is a much better answer than "approved."

Version your policy too. Rules for grace periods and acceptable damage will change, and you need to know which rule applied to each historical decision.

Putting It Together

ID Analyzer gives you the signals this framework depends on. It provides OCR across 3,000+ document formats from 190+ countries, MRZ and barcode reading for cross-checks, document authentication and anti-forgery analysis, and biometric face match with liveness. Your application layer then applies the accept, flag or rescan logic that fits your regulatory obligations.

検証を始める

最初の ID を検証する準備はできましたか?

ご登録で無料のテストクレジットを付与。カード登録は不要です。

  • クレジットカード不要

  • 登録で無料トライアルクレジット進呈