🎉 Unlimited Free KYC - Forever!!

Identity Verification
Facial Recognition
Facial Similarity Report

Facial Similarity Report

Our facial recognition process compares the live image or video of the user with the portrait extracted from their identity document. This comparison ensures that the person presenting the document is indeed the document's owner. Below is a detailed explanation of the fields you can expect in a facial similarity report.

Sample Facial Similarity Report

Report Structure

The facial similarity report returns a JSON object with a root-level face object containing all verification results.

Core Response Fields

interface FaceVerificationResponse {
  face: {
    status: 'Approved' | 'Rejected' | 'In Review' | 'Not Finished';
    face_match_status: 'Approved' | 'Rejected' | 'In Review' | 'Not Finished';
    liveness_status: 'Approved' | 'Rejected' | 'In Review' | 'Not Finished';
    face_match_similarity: number;  // Range: 0-100
    liveness_confidence: number;    // Range: 0-100
    source_image: string;          // Temporary URL
    target_image: string;          // Temporary URL
    video_url: string;            // Temporary URL
  }
}

Response Fields

Status Fields

  • status: Overall verification status combining face match and liveness results
  • face_match_status: Result of facial comparison between live image and document photo
  • liveness_status: Result of anti-spoofing verification

Confidence Scores

  • face_match_similarity: Normalized similarity score between live image and document photo (Range: 0-100)
  • liveness_confidence: Anti-spoofing confidence score indicating probability of a real person (Range: 0-100)

Media References

  • source_image: HTTPS URL to the portrait extracted from ID document (JPG, max 1920x1080)
  • target_image: HTTPS URL to the best frame from verification (JPG, min 720p)
  • video_url: HTTPS URL to raw verification video (MP4, max 15 seconds)

Sample Response

{
  "face": {
    "status": "In Review",
    "face_match_status": "Approved",
    "liveness_status": "In Review",
    "face_match_similarity": 95.43,
    "liveness_confidence": 65.0,
    "source_image": "https://example.com/source-image.jpg",
    "target_image": "https://example.com/target-image.jpg",
    "video_url": "https://example.com/video.mp4"
  }
}

Understanding the Results

Face Match Status

  • "Approved" indicates a successful match between the live image and the document portrait.
  • "Rejected" suggests a significant discrepancy, potentially indicating fraud.
  • "In Review" may require manual review or additional verification steps.

Liveness Status

  • "Approved" confirms that the system detected a live person during the verification process.
  • "Rejected" suggests potential use of a static image, video, or mask.
  • "In Review" may indicate inconclusive results requiring further review.

Face Match Similarity

  • Scores typically range from 0 to 100, with higher scores indicating greater similarity.
  • A score of 95.17 (as in the example) suggests a very high degree of similarity.
  • The threshold for approval can be customized based on your risk tolerance.

Liveness Confidence

  • Similar to the similarity score, this ranges from 0 to 100.
  • A score of 96.0 (as in the example) indicates high confidence in the liveness of the presented face.
  • This helps prevent spoofing attempts using photos or videos.

For complete API specifications and additional fields, refer to our API Reference.

Security Considerations

  • URLs expire after 60 minutes
  • We recommend:
    • Fetching images immediately if needed
    • Storing only verification results
    • Implementing proper access controls
    • Regular audit of stored data