Deduplication as a Service: Overcoming the myth of ABIS Proprietary Solutions

Deduplication of Identities Using Similarity Search in a Scalable Vector Database

“…According to a report from Transparency Market Research 1 , the market value of biometrics systems is expected to increase 3.7 times from 2021 to 2031, and a key factor in its growth is the demand of biometric technologies by government and enterprise systems for authentication and identification purposes. However, current commercial biometric solutions are cost prohibitive because there are only a few vendors that offer a complete and trustworthy deduplication solution, and hence unaffordable by most underdeveloped/developing nations [1].…”

“…A free AFIS (Automated Fingerprint Identification System) solution used to manage the identification of citizens of Rondônia, Brazil proved to be unfeasible because of high querying time, and low scalability [1]. Another open-source solution OpenABIS, launched in 2019 2 and cited as the first open source ABIS, employs Dlib [2] based face recognition algorithm.…”

“…Other factors such as the birth rate, and mortality rate also influence the growth of the system. The case of brazilian enrollment system, perhaps can be cited as an example here, where an on-premise solution was implemented with a high processing power requirements, and after the initial enrollment of the population, the compute requirements declined by 36 times within a month [1] leaving the system idle, and wasteful. It would take approximately ten more years for the data to grow to the point where it can fully utilize the resources, leading to the wastage of key national resources and a non-scalable nature of the system.…”