A clear passport or identity-card image is not proof of a trustworthy application. The document may be altered, expired, reported stolen, or presented by someone other than its legitimate holder.
Digital verification therefore needs to establish three things: whether the document is authentic, whether it is acceptable and valid, and whether it belongs to the applicant.
Face++ OCR, face comparison, and liveness detection can support the data-extraction and holder-verification layers within this broader process.
1. Authenticity, Validity, and Holder Verification
These checks answer different questions:
- Authenticity: Was the evidence genuinely issued, without unauthorized alteration?
- Validity: Is it current and acceptable for this verification purpose?
- Holder verification: Is the applicant the person associated with the evidence?
An authentic document can be expired. A valid document can be stolen. A matching face cannot independently establish that the document was genuinely issued.
NIST similarly distinguishes evidence validation—checking authenticity, accuracy, and validity—from verification of the applicant’s relationship to that evidence. NIST Identity Proofing Overview
2. Capture Quality and Document Recognition
Validation starts with usable evidence. Blur, glare, cropping, and low resolution can obscure security features or produce extraction errors.
A capture workflow can check:
- Visibility of required document edges and surfaces
- Text and portrait clarity
- Reflections and obstruction
- Signs of screen or print recapture
- Document type, issuing country, and version
Recognizing the correct document version matters because layouts and security features change over time.
Poor capture quality should normally trigger a guided recapture, not an automatic fraud accusation. Face++ OCR can contribute to extracting visible information, but supported documents and languages should be confirmed for the selected integration.
3. OCR and Internal Data Consistency
OCR converts visible text into structured fields such as name, date of birth, document number, and expiry date.
The verification system can then compare:
- Printed fields with machine-readable-zone or barcode data
- Information across the front and back
- Dates against expected formats and logical relationships
- Document numbers against applicable formatting rules
- Extracted details with information supplied by the applicant
Where applicable, machine-readable check digits help identify inconsistent data. However, a correct check digit does not prove that an issuer created the document.
Face++ text-recognition capabilities support extraction; authenticity requires additional evidence beyond readable text. Face++ Capabilities
4. Visual and Security-Feature Examination
Document analysis evaluates whether the submission is consistent with the claimed document type and version.
Relevant indicators include typography, field placement, background patterns, portrait boundaries, and evidence of digital compositing or physical alteration.
Security-feature checks must match the capture method. An ordinary photograph cannot reliably establish every ultraviolet, tactile, or angle-dependent feature. Some checks require controlled lighting, multiple views, specialized equipment, or chip access.
A convincing appearance is therefore supporting evidence—not a guarantee of authenticity.
5. Chip Authentication and Trusted-Source Checks
For supported electronic passports, chip authentication can provide stronger evidence of issuer origin and data integrity.
The system must verify digital signatures and the relevant certificate chain. Simply reading information through NFC is not equivalent to authenticating it. ICAO explains that electronic travel-document validation uses document-signer certificates, issuing-country trust certificates, and certificate-revocation information. ICAO PKD FAQ
Chip-signature validation does not, by itself, prove that the presenter is the holder or that the document has not been reported stolen.
Where authorized access exists, issuer or trusted-source checks can provide additional confirmation of identity attributes or document status. Availability varies; not every document has an accessible verification database.
6. Expiry, Status, and Acceptance Rules
A genuine document must still satisfy the platform’s acceptance policy and applicable requirements.
Checks may include:
- Expiry date and permitted validity window
- Accepted document types and issuing jurisdictions
- Required identity attributes
- Cancellation or lost-and-stolen status, where accessible
- Whether additional evidence is necessary
A database outage should not be recorded as a confirmed invalid document. Similarly, an unsupported document should be distinguished from a fraudulent one.
Keeping these outcomes separate supports clearer customer guidance and more accurate risk decisions.
7. Linking the Document to the Applicant
Document validation does not prevent someone from presenting another person’s genuine ID.
Face++ 1:1 face comparison can compare the document portrait with a current facial capture and return a confidence score for threshold-based assessment. The score supports a similarity decision; it does not authenticate the document itself. Face++ Face Comparing
Liveness detection addresses a separate concern: whether the facial sample represents a live person rather than a presentation such as a photograph, replayed video, or mask. Face++ offers liveness capabilities for this anti-spoofing layer. Face++ Liveness Detection Overview
Capture-integrity controls should separately address manipulated input paths. Neither a face match nor a liveness result should override unresolved document-authenticity concerns.
8. Combining Evidence into a Decision
A layered workflow combines document findings, data consistency, available chip or source checks, Face++ biometric results, and session context.
Possible outcomes include:
- Approve: Required checks provide sufficient confidence.
- Recapture: Evidence quality prevents reliable assessment.
- Review: Conflicting or incomplete findings require investigation.
- Reject: Confirmed fraud or unmet acceptance requirements justify rejection.
Record the checks performed, unavailable checks, relevant results, policy version, and decision reason. Protect retained identity data through appropriate access controls and retention limits.
9. Frequently Asked Questions
Q1. Does successful OCR prove an identity document is genuine?
No. OCR can accurately extract information from an altered or counterfeit document.
Q2. Can a genuine document fail digital verification?
Yes. It may be expired, unsupported, poorly captured, or presented by someone other than its holder.
Q3. How does Face++ support identity document validation?
Face++ OCR supports data extraction, while face comparison and liveness detection support applicant verification. These capabilities complement document-authenticity analysis and available trusted-source checks.
Q4. Does a matching face prove the document is authentic?
No. An attacker’s face could match a substituted portrait. Document authenticity and holder verification must be evaluated separately.
10. Build Confidence Across the Whole Verification Process
Reliable digital verification requires more than readable fields or a convincing document image.
Combining document examination, consistency checks, supported chip authentication, and Face++ face comparison and liveness detection helps businesses assess both the evidence and its presenter. Clear recapture and review paths protect legitimate applicants when the available evidence remains uncertain.
