Document Liveness

Verifies that an identity document is authentic and not a photocopy, printout, or digital recreation.

How It Works

Valydar analyses 6 signals from the document image:

1. Background uniformity — real documents have printed backgrounds
2. Lighting gradient — natural light falls unevenly on real documents
3. Edge softness — real documents have specific edge characteristics
4. Moiré detection — prints and screens create moiré patterns
5. Colour quantisation — real documents have continuous colour tones
6. Perspective hint — geometric analysis of document shape

Run Document Liveness

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

Result

{
  "passed": true,
  "score": 0.93,
  "checks": [
    {
      "name": "background",
      "passed": true,
      "score": 0.95,
      "reason": "uniform background detected"
    },
    {
      "name": "lighting_gradient",
      "passed": true,
      "score": 0.91,
      "reason": "natural lighting gradient"
    },
    {
      "name": "edge_softness",
      "passed": true,
      "score": 0.89,
      "reason": "edges are soft, not printed"
    },
    {
      "name": "moire_pattern",
      "passed": true,
      "score": 0.96,
      "reason": "no moiré pattern detected"
    },
    {
      "name": "colour_quantisation",
      "passed": true,
      "score": 0.92,
      "reason": "natural colour distribution"
    },
    {
      "name": "perspective_hint",
      "passed": true,
      "score": 0.94,
      "reason": "correct perspective geometry"
    }
  ]
}

Score Thresholds

Score	Meaning
> 0.8	Likely authentic document
0.5 - 0.8	Borderline — may be a good photocopy
< 0.5	Likely fake or copy

Detected Spoof Types

Attack Type	Detected
Photocopy	✅
Printout	✅
Screen capture	✅
Good quality scan	⚠️ Partial
Tampered document	❌ (use document check for data validation)