Skip to main content
The Document Classification module identifies the type of identity document a user has submitted by analysing captured images. It determines the document type, for example, whether it’s a passport, driving licence, or national ID card, and where it was issued.

Document Classification V2

Document Classification will process both single and double-sided Identity Documents. The output includes a Classification status, Document Type, Document Issue Year, Country Code, State Name (if available) and other Capture-level checks. The complete classification details are also included in the Document object under Subject.

Capabilities

The module returns a structured set of capabilities grouped into three areas: classification result, document identification, and image capture quality checks. Each capability returns a single value that downstream evaluation logic can act on.

Document classification result

The overall outcome of the classification attempt, indicating whether the module was able to identify the document type.

Document type

The type of identity document identified during classification (for example Passport, Identification Card, or Driving License). The value determines which verification rules and reference templates are applied by downstream modules such as Document Authentication.

Document country code

This capability returns the ISO 3166-1 alpha-3 country code of the country that issued the document (e.g. GBR for the United Kingdom, USA for the United States, AUS for Australia). The module supports documents from over 190 countries. If the issuing country cannot be determined, the value defaults to Unknown. The full list of supported country codes follows the ISO 3166-1 alpha-3 standard. Common examples include:

Country name

This capability returns the full name of the issuing country as a human-readable string (e.g. “United Kingdom”, “United States of America”, “Australia”). This value corresponds to the country code returned by the document country code capability. If the country cannot be determined, the value defaults to Unknown.

State name

This capability returns the name of the state, province, or territory that issued the document, where applicable. This is particularly relevant for countries where identity documents are issued at the sub-national level, such as driving licences in the United States and Australia. The module supports the following regions:
  • United States: All 50 states and the territories of Guam and American Samoa (via the country code, not the state name field).
  • Australia: All states and territories, including New South Wales, Queensland, Victoria, South Australia, Western Australia, Tasmania, Northern Territory, and Australian Capital Territory.
  • United Kingdom and Crown Dependencies: Northern Ireland, Wales, Guernsey, Jersey, and Isle of Man.
  • Caribbean: Curaçao.
If the document was not issued at the state or territory level, or if the state cannot be determined, the value defaults to Unknown.

Image capture quality checks

The module performs a series of image quality assessments on both the front and back sides of the submitted document. These checks evaluate whether the captured images are of sufficient quality for accurate classification and downstream processing. Poor image quality is one of the most common causes of classification failure. Glare from reflective document surfaces, motion blur from camera movement, low resolution from distant capture, washed-out colour from poor lighting, and cropped edges from incorrect framing can all prevent the module from identifying the document. Each capture quality check returns one of the following values: The following checks are performed:
  • Low resolution check:
Evaluates whether the captured image has sufficient pixel density for the module to accurately identify document features. Low-resolution images may result from the camera being too far from the document or from a device with a low-quality camera sensor. Insufficient resolution makes it difficult to read text, identify security features, and classify the document accurately.
  • Glare check:
Detects reflective glare on the document surface. Glare typically appears as bright white spots or washed-out areas caused by overhead lighting, camera flash, or the reflective laminate coating found on many modern identity documents. Glare can obscure printed text, photographs, and security features, making the document unreadable to the classification engine.
  • Blur check:
Detects motion blur or focus blur in the captured image. Blur occurs when the camera or the document moves during capture, or when the camera fails to focus correctly on the document surface. Blurred images reduce the clarity of text, edges, and fine details that the classification engine relies on to identify document type.
  • Colour check:
Determines whether the image was captured in full colour. Some document features, such as colour-coded regions, flag emblems, and security holograms, require colour information for accurate classification. A greyscale or black-and-white image, for example, from a photocopy may lack the visual information needed to classify the document correctly. This check can also help detect photocopied documents, which are commonly used in fraud attempts.
  • Full document in view check:
Verifies that the entire document is visible within the captured image frame. If any edges of the document are cropped or cut off, the module may be unable to identify the document type, read border-area text (such as MRZ data on passports), or detect edge-based security features. This check ensures that the user positioned the document correctly within the camera’s field of view during capture.

Default outcomes

The module is pre-configured with the following default outcomes, which can be used in evaluation and routing logic:

Input payload

The following is a sample payload used to submit document images to the Document Classification module for processing.
JSON

Sample response

The following is a sample response for a successfully classified document.
JSON

Document classification V1

Capabilities

Document classification result

The overall outcome of the classification attempt, indicating whether the module was able to identify the document type.

Document type

The type of identity document identified during classification (for example Passport, National Identification Card, or Driving License). The value determines which verification rules and reference templates are applied by downstream modules such as Document Authentication.

Document sub type

Returns the Document sub-type - a Driving Licence sub-type could include e.g. “Provisional Driving License” or “Learner Driving License”

State name

Returns the state name if applicable.

Document country code

Returns the country code in ISO3 format, eg. GBR.

Country name

Returns the full country name

Front side low resolution check

Assesses whether the front side image has sufficient pixel density to identify document features accurately. Low resolution, from the camera being too far away or a low-quality sensor, makes text, security features, and document type harder to identify.

Back side low resolution check

Assesses whether the back side image has sufficient pixel density to identify document features accurately. Low resolution, from the camera being too far away or a low-quality sensor, makes text, security features, and document type harder to identify.

Front side glare check

Detects reflective glare on the front side of the document, such as bright spots or washed-out areas from overhead lighting, camera flash, or reflective document laminate, which can obscure printed text, photographs, and security features.

Back side glare check

Detects reflective glare on the back side of the document, such as bright spots or washed-out areas from overhead lighting, camera flash, or reflective document laminate, which can obscure printed text, photographs, and security features.

Front side blur check

Detects motion or focus blur on the front side image, caused by camera or document movement during capture or incorrect focus, which reduces the clarity of text, edges, and fine detail used for classification.

Back side blur check

Detects motion or focus blur on the back side image, caused by camera or document movement during capture or incorrect focus, which reduces the clarity of text, edges, and fine detail used for classification.

Front side colour check

Checks whether the front side image was captured in full colour. Colour-coded regions, flag emblems, and security holograms require colour information; a greyscale image (for example a photocopy) may lack the detail needed to classify the document and can indicate a photocopied document.

Back side colour check

Checks whether the back side image was captured in full colour. Colour-coded regions, flag emblems, and security holograms require colour information; a greyscale image (for example a photocopy) may lack the detail needed to classify the document and can indicate a photocopied document.

Front side full document in view check

Verifies that the entire front side of the document is visible within the captured image frame. Cropped or cut-off edges can prevent identifying the document type, reading border-area text such as passport MRZ data, or detecting edge-based security features.

Back side full document in view check

Verifies that the entire back side of the document is visible within the captured image frame. Cropped or cut-off edges can prevent identifying the document type, reading border-area text such as passport MRZ data, or detecting edge-based security features.

Front side photo substitution check

Returns Front Side Photo Substitution Check result

Back side photo substitution check

Returns Back Side Photo Substitution Check result

Front side digital tamper check

Returns Front Side Digital Tamper Check result

Back side digital tamper check

Returns Back Side Digital Tamper Check result

Front side document liveness check

Returns Front Side Document Liveness Check result

Back side document liveness check

Returns Back Side Document Liveness Check result

Input payload

The following is a sample payload used to submit data to the Document classification V1 module for processing.
JSON

Sample response

The following is a sample response returned by the module.
JSON