Technologies¶
Document Type Identification¶
Document Type Identification helps detect forged or invalid documents. This technology utilizes artificial intelligence and machine learning to ascertain the presented document’s type and legitimacy by comparing it to a reference database of verified templates. Any document lacking proper formatting or layout can be singled out for further inspection.
Document Type Identification can be performed using the document image from our extensive database, and it’s a straightforward and intuitive process. First, the user scans a document using their mobile device or captures it via the web. The scan is then evaluated, the document type is identified, and its key attributes are compared to corresponding template images. All in all, this process takes just a second, regardless of the identity document image submitted from any device, be it a mobile phone or laptop, using the id3 Document SDK.
See Supported documents for the list of currently supported documents.
MRZ Reading¶
The Machine Readable Zone (MRZ) comprises either two or three lines of encoded information, typically located at the lower portion of an identification document. It is where crucial personal details of the document holder are encoded in a standardized format, allowing for swift machine-based reading and verification.
The introduction of MRZ technology in the 1980s aimed to expedite ID verification processes at borders and airports. Thanks to its precision and speed, various organizations and services quickly adopted this technology. Today, machine-readable zones are a feature of various documents, collectively known as Machine Readable Travel Documents (MRTD).
The following MRZ types are currently supported by this SDK:
Document TD1 (ICAO 9303-5) - 3x30 lines MRZ.
Document TD2 (ICAO 9303-6) - 2x36 lines MRZ.
Document TD3 (Passports)(ICAO 9303-4) - 2x44 lines MRZ.
Driving licence (ISO-18013) - 1x30 line MRZ.
Visa format A (ICAO 9303-7) - 2x44 lines MRZ.
Visa format B (ICAO 9303-7) - 2x36 lines MRZ.
Other MRZ Type, with a standard layout but unknown content.
Old french ID card (PRADO FRA-BO-02002). 2x36 lines MRZ.
Optical Character Recognition¶
Optical Character Recognition (OCR) is the technology that transforms typed or printed text into machine-readable data.
Once this conversion is complete, the text can be easily edited, searched, displayed, or stored electronically. OCR has traditionally been employed for tasks such as data entry and the conversion of printed archival documents (such as bank statements and insurance records) into a digital format.
Advanced OCR systems employ a variety of pre-processing techniques to reduce errors and enhance character recognition. These techniques can include de-skewing, de-speckling, and normalization of the images processed by OCR.
The id3 Document SDK leverages pattern recognition, artificial intelligence, and computer vision, greatly expanding its capabilities and applications. Today, it has become a standard technology used across a wide range of industries.”