Face Match

Compares a selfie to the photo on the uploaded document to confirm the same person.

How It Works

1. User uploads a document (with photo)
2. User uploads a selfie
3. Valydar compares the two faces using colour histograms, texture analysis, and structural similarity

Upload a Selfie

curl -X POST https://api.dev.valydar.com/verifications/{id}/selfie \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "file=@selfie.jpg"

Run Face Match

curl -X POST https://api.dev.valydar.com/verifications/{id}/face-match \
  -H "Authorization: Bearer YOUR_API_KEY"

Result

{
  "passed": true,
  "confidence": 0.87,
  "threshold": 0.45,
  "algorithm": "combined",
  "document_face_detected": true,
  "selfie_face_detected": true
}

Score Thresholds

Score	Meaning
> 0.7	High confidence match
0.45 - 0.7	Match (above threshold)
< 0.45	No match (below threshold)

Tips

• Ensure good lighting for the selfie
• Face should be clearly visible and centred
• Avoid sunglasses or heavy filters