Identity Check is an endpoint built to facilitate KYC (Know Your Customer) processes in digital products.

On a digital product, the user uploads his/ her identification card, takes a live picture with their device's camera and fills out a form with their details. The live image, the identification card and their name is sent to this endpoint and the endpoint verifies using image recognition that the face in the identity card matches that which is in the live image. After which, using optical character recognition, verification is made that the name passed in from the form matches that on the identity card.

Follow these steps to set up and run the service locally :

Clone the repository :

git clone https://github.com/Urias-T/identity_check.git

Navigate to the project directory :

cd identity_check

Create a virtual environment and activate it :

python -m venv venv
venv/Scripts/activate

Install the required dependencies in the virtual environment :

pip install -r requirements.txt

Launch the copilot service locally :

python main.py

That's it! The service is now up and running on localhost:8080. 🤗

To run this as a docker container:

Clone the repository :

git clone https://github.com/Urias-T/identity_check.git

Navigate to the project directory :

cd identity_check

Build the Docker image:

docker build -t identity_check .

Take note to include the dot at the end of the docker command above

Run the Docker container:

docker run -p 8080:8080 -d identity_check

Congratulations, your service will be running on localhost:8080 🎉

Endpoint:

• Method: POST
• Route: /verify

Request Parameters:

• image1: Live image file (Required)
• image2: ID image file (Required)
• name: Name associated with the ID (Required)

Supported File Types:

• Supported file types for image uploads: jpg, jpeg, png

Example Usage:

{
  "image1": "live_image.jpg",
  "image2": "id_image.jpg",
  "name": "John Doe"
}

Response Codes:

• 200: Successful verification.
• 400: Bad request - Missing image file(s), no selected file, no face detected in image(s), or invalid file format.
• 401: Unauthorized - Face mismatch or name mismatch.
• 500: Internal Server Error - An unexpected error occurred during the verification process.

Response Body:

1. Success Response (HTTP 200):

  {
    "status": "Success",
    "message": "Verified",
    "details": {
        "face_match": true,
        "name_match": true
    }
}

2. Bad Request (HTTP 400):

• Missing Image File(s):

    {
        "status": "Error",
        "message": "Invalid Request",
        "details": {
             "error": "Missing image file(s)!"
         }
    }

• No Selected File(s):

      {
          "status": "Error",
          "message": "Invalid Request",
           "details": {
               "error": "No selected file(s)"
            }
     }

• No face detected in image(s):

      {
          "status": "Error",
          "message": "An error occured",
           "details": {
               "error": "There was no face in at least one of the images."
            }
     }

• Invalid file format:

      {
          "status": "Error",
          "message": "Invalid Request",
           "details": {
               "error": "Invalid file format(s)"
            }
     }

3. Failure Response (HTTP 401):

• Face Mismatch:

    {
        "status": "Failure",
        "message": "Verification failed. Live image does not match ID image.",
        "details": {
            "face_match": false,
            "name_match": false
        }
    }

• Name Mismatch:

    {
        "status": "Failure",
        "message": "Verification failed. Provided name doesn't match that on ID.",
        "details": {
            "face_match": true,
            "name_match": false
        }
    }

4. Error Response (HTTP 500):

{
    "status": "Error",
    "message": "An error occurred.",
    "details": {
        "error": "<error_message>"
    }
}

Error Conditions:

• If any required parameter is missing, the endpoint will return a 400 error with an appropriate error message.
• If the selected file for either image is empty or the file format is not supported, a 400 error will be returned.
• If there is no image detected in any of the images, the endpoint would return a 400 error with an appropriate error message.
• In case of face or name mismatch, the endpoint returns a 401 error with details indicating the reason for failure.
• If an unexpected error occurs during the verification process, a 500 error will be returned with details of the error.

If you want to contribute to this project, please open an issue and submit a pull request.

This project is made available under the MIT License.