There are two ready-made database solutions for Document Reader SDK:
Full database includes all the documents of all the countries listed in the Documents List.
FullAuth type includes all the documents of all the countries plus all the security checks that are indicated in the Security Features column of the Documents List.
If you do not need the full database, you can shrink your application size by excluding some of the excess items.
graph LR A(How to shrink<br>the database?) --> B(By countries); A --> C(By document types); A --> D(By countries &<br>document types); A --> E(By types of<br>certain documents); A --> F(By document types<br>with RFID chip<br>embedded); A --> G(By languages); A --> H(By security checks);
On this page, several ways to shrink the database are collected. Please contact your account manager or the Support Portal to get a configured custom database.
You can exclude unnecessary countries and leave only those you need.
You can leave only the types of documents you work with: passports, driving licenses, IDs, visas, etc.
You can leave included all the documents from the selected country and add only chosen types of documents of other countries. For example, all the types of documents of Germany plus passports of all the other countries.
In some countries, some types of documents may have several subtypes, for example, type of passports by the issuing year. You can include to your database only certain types of certain documents.
You can leave included only documents that have an RFID chip embedded.
By default, the SDK can read text in languages that use the Latin alphabet. Besides, the
Full database includes three neural networks for other languages. You can exclude all of them or leave one or two.
The neural networks recognize the following groups of languages:
- Type 1: Chinese, Japanese, Korean.
- Type 2: Languages of Algeria, Bahrain, Egypt, Iraq, Jordan, Lebanon, Arabic World, Libya, Morocco, Oman, Qatar, Saudi Arabia, Arabic-Syria, Tunisia, U.A.E, Yemen, Persian, Urdu, Pashto, Sindhi.
- Type 3: Lao, Thai, Sinhala, Bengali-Bangladesh, Bengali-India, Tamil, Hindi, Kashmiri, Konkani, Marathi, Punjabi, Sanskrit, Sindhi, Amharic.
The security checks can be excluded or set selectively. There are two groups of security checks available:
- For authenticity checks in ultraviolet (spectrum) light, applicable for Regula Mobile document authenticators.
- For Visible/IR patterns authenticity checks.