Document recognition

You can extract standard fields and recognize the text of templates and documents, such as passports, driver's licenses, vehicle registration certificates, and license plate numbers.

Recognition modelsRecognition models

Specify the model you need in the model field in your request.

Models for recognizing common documents:

• passport: Passport data page.
• driver-license-front: Driver license, front side.
• driver-license-back: Driver license, back side.
• vehicle-registration-front: Vehicle registration certificate, front side.
• vehicle-registration-back: Vehicle registration certificate, back side.
• license-plates: All license plate numbers in the image.

Countries whose documents can be recognizedCountries whose documents can be recognized

List of returned fieldsList of returned fields

Depending on the model, the list of fields in the response may vary:

passport model:

• name: Name.
• middle_name: Middle name.
• surname: Last name.
• gender: Gender.
• citizenship: Citizenship.
• birth_date: Date of birth.
• birth_place: Place of birth.
• number: Passport number.
• issued_by: Issued by.
• issue_date: Issue date.
• subdivision: Department code.
• expiration_date: Date of expiration.

driver-license-front model:

• name: Name.
• middle_name: Middle name.
• surname: Last name.
• number: Driver license number.
• birth_date: Date of birth.
• issue_date: Issue date.
• expiration_date: Date of expiration.

driver-license-back model:

• experience_from: Driving experience since (year).
• number: Driver license number.
• issue_date: Issue date.
• expiration_date: Date of expiration.
• prev_number: Driver's previous license number.

vehicle-registration-front model:

• stsfront_car_number: License plate number.
• stsfront_vin_number: Vehicle identification number (VIN).
• stsfront_car_brand: Vehicle brand.
• stsfront_car_model: Vehicle model.
• stsfront_car_year: Vehicle manufacture year.
• stsfront_car_chassis_number: Chassis number.
• stsfront_car_trailer_number: Vehicle body, compartment, or trailer number.
• stsfront_car_color: Vehicle color.
• stsfront_sts_number: Vehicle registration certificate (VRC) number.

vehicle-registration-back model:
* stsback_car_owner: Vehicle owner's full name or legal entity name.
* stsback_sts_number: Vehicle registration certificate (VRC) number.

ExampleExample

Recognition requestRecognition request

body.json file:

{
  "mimeType": "<data_type>",
  "languageCodes": ["*"],
  "model": "<recognition_model>",
  "content": "<image_for_recognition>"
}

Where:

• mimeType: MIME type of the data to transmit.
• languageCodes: Text language. To detect the text language automatically, specify *.
• model: Recognition model.
• content: Base64-encoded image.

Response to a requestResponse to a request

The response of the passport, driver-license-front, driver-license-back, vehicle-registration-front, and vehicle-registration-back models will contain the entities array.

The response of the license-plates model does not contain the entities array. This model recognizes all license plate numbers in the image and does not recognize any other text. In this case, the completeness and accuracy the license plate numbers are recognized for this model are much higher than for the general OCR model. The recognition results are shown in the standard text_detection response.

Here is an example of the entities array returned in the service response for the passport model:

{  "entities": [
   {
    "name": "name",
    "text": "Sonya"
   },
   {
    "name": "middle_name",
    "text": "Anna"
   },
   {
    "name": "surname",
    "text": "Ray"
   },
   {
    "name": "gender",
    "text": "female"
   },
   {
    "name": "citizenship",
    "text": "usa"
   },
   {
    "name": "birth_date",
    "text": "12.05.1978"
   },
   {
    "name": "birth_place",
    "text": "Ap #762-7228 Dis Rd.Austin KY 50350."
   },
   {
    "name": "number",
    "text": "0702084625"
   },
   {
    "name": "issued_by",
    "text": "washington state department of licensing."
   },
   {
    "name": "issue_date",
    "text": "10.05.2011"
   },
   {
    "name": "expiration_date",
    "text": "-"
   }
  ]}

See alsoSee also

• List of supported languages and models
• Known issues in the current Vision OCR version
• How to recognize text in an image