Computer Vision News - October 2016

patterns and traits. In that way, if any similar patterns emerge on future documents in the future, the app will give an additional alert to their clients. Then clients can double check and focus more resources on the individuals that pose a higher risk. Kassai told us that usually other identity verification providers work manually. Each time an image comes on their computer screen for the data processors, they compare these templates manually. Instead, Onfido verifies these documents using an automated process with six layers of verification . The first layer is Optical Character Recognition . When a passport comes through the system, the latter uses OCR to extract all of the characters and verify whether they are correct. A good thing with passports and many other identity documents is that you can do a parity check. Usually passports contain a standardized machine readable zone at the bottom which includes a parity check, so you can tell if the extracted passport number expiration date is consistent with the passport machine readable zone. That is like an internal consistency accuracy check. Whether or not the app checks with authority databases or with an autonomous system depends very much on the country. There are two types of databases you can check. The first type of database is one which allows you to check whether the government issued that document or not. In India for example, there is a Unique Identification Authority (UIDAI) and Onfido checks the Indian identity card (called Aadhaar) by extracting its number and sending it to the government’s API database. Then, they receive a confirmation as to whether the name, date of birth, and Aadhaar card number was issued by them or not. The process is the same in the 15 states across the US which allow to go through it via the Department of Motor Vehicles (DMV). They also do it in the UK and in Brazil (with the CPF numbers). On the other hand, Onfido does not need to check passports with anyone elsewhere in the world, because consistency and parity checks are built in the machine readable zone. This is one type of database research which would validate whether or not the issuing body had issued that document. The other databases that they search are police databases, which depend on each different country where Onfido establishes relationships with law enforcement authorities. Every time police raids a warehouse with thousands of fake passports, they extract the numbers and communicate them to Onfido. When the latter finds Computer Vision News Application 21 Application Husayn Kassai, Onfido’s CEO and co-founder

RkJQdWJsaXNoZXIy NTc3NzU=