Face Match

Decentro provides capabilities to platforms to easily verify the identity of the customers and help identify if the customer's photographic identity is genuine like matching a selfie with that from a KYC document or verifying if the customer is actually in front of the app.

Face Match


Face match technology, also known as face recognition is used to compare one image having a face to one or more images having the face of the person. It uses deep learning, face match technology is capable to identify if the images with faces belong to the same person, and whether they are matching with a minimum threshold value or not.

Use Cases of Face Match

Face match technology, which can compare, verify and recognize faces within a fraction of seconds can be used for various purposes. Few of them are -

  • KYC (Know Your Customer)
  • Prevent Identity Fraud
  • Surveillance

Documents Supported

We support 3 formats as of now for Face Match API. They are -

  • JPEG
  • JPG
  • PNG



We will work on other formats in near future.

Face Match: Parameters and Description

Threshold - The confidence at which you want to verify the identity. The minimum threshold by default is 0.7. It can be modified to a max threshold of 0.95.

Image Size - The maximum size limit of image upload and URL is 6MB

Note - There are 3 ways to verify the images with the Face Match service:

  • Upload both images from the local device.
  • Provide URL to both the images.
  • Go hybrid, upload the first image, and then provide the URL to the second image or vice versa.
curl --location --request POST 'https://in.staging.decentro.tech/v2/kyc/forensics/face_match' \
--header 'client_id: <your_client_id>' \
--header 'client_secret: <your_client_secret>' \
--header 'module_secret: <your_module_secret>' \
--form 'reference_id="eefa-abs1-0001-8915"' \
--form 'consent="Y"' \
--form 'consent_purpose="for bank account purpose only"' \
--form 'image1=@"/path/to/file"' \
--form 'image2=@"/path/to/file"'
    "status": "SUCCESS",
    "responseCode": "S00000",
    "message": "Face scan completed successfully.",
    "data": {
        "status": "SUCCESS",
        "match": "95.00%"
    "responseKey": "success_face_scan"