Technocrat Consult and IT Systems Limited (Technocrat IT) has spent the last 5 years primarily focused in research and development of biometrics, developing several systems and processes for biometric implementation. The unique advantage of Technocrat IT is our advanced knowledge and experience in Hardware and software giving us the unique vantage point of developing an End-to-End solution, guarantying success.
The lack of understanding of the numerous challenges faced with biometrics and how to use the combined effect of hardware and software to combat it has been a major issue with biometric projects. The infrastructural challenges of Nigeria also adds to the biometric challenge. Technocrat IT has made it our primary business to develop expertise in biometric. The success of this is evident in the end result of our 5 years of hard work, the ownership of the copyright-PATENT “NG/P/2010/283” for “Portable Telecommunication (Mobile Client) Device used in Biometric Identification System”. The Topic of the Patent and its description is highly Technical. However, we would make an attempt below to explain in simple layman terms the coverage of our patent and its benefit to Biometric identification as regards on the Field capture. The high level summary of our patent are three related areas:
1. The use of a Portable Telecommunication device for on the field data capture for effective biometric identification system. This is now popularly known as Direct Data Capture Machine (DDCM). This must have FIPS 201 compliant finger scanner/reader i.e. minimum 500dpi and high resolution camera. Minimum 2 Mega pixel.
2. Validate all captured biometrics data. Ensure correct data is entered. e.g, All captured data has photograph of a real person, has finger print of a live/real person.
3. Identify Duplicate registrations/information using the biometric data captured using the DDCM.
The identified duplicate can then be processed as deemed fit based on previously defined processes unique to the organization using it. It might be to prevent duplicate (National ID card and voters register), to simply identify matching duplicate (Criminal database) or merge duplicates (SIM registration).
The three key areas were missing in the previous exercises using technology and biometrics (National ID card, 2006 voters register), and hence the massive failure. Our patent and process(es) is designed to fix this with description of the technical process(es) to follow to achieve this.
This patent covers the two major processes for achieving a successful Biometric registration or identification project. The patent covers the Use of Portable Biometric devices in Biometric identification systems (a registration process or actual identification process are all part of an identification system).The patent covers the use of 1 or more of the two unique preposition in this Patent which are:
1. Use of DISTRIBUTED BIOMETRICS: This is ensuring both Front end machines and the back end infrastructure server all cooperate in the Biometrics processing to achieve an effective result. Some critical components of these cooperation are as below:
a. BIOMETRIC DATA CAPTURE: Information and Biometric data capture with the DDCM (Direct Data capture machines, which should have finger print scanner [preferably FIPS specification of minimum 500dpi] and high resolution camera. This machines can be hand held devices, laptops, Notebooks, Mobile Phones or any other portable Telecommunication device). In cases of registration, the DDCM is used to capture details including fingerprint(s) and facial ('passport') photograph (Mug-shot) of the registrant.
b. FRONT END BIOMETRIC PROCESSING using the DDCM: The DDCM should process the biometric in one or multiple of the following ways
i. Check that the finger print is present and valid real person finger print.
ii. Check that the Photo captured is present and valid of appropriate dimensions of a human face
iii. Check for duplicates or previous entries for this individual on the DDCM
iv. Encryption of the Biometric data for easier processing and security.
c. BACK END BIOMETRIC PROCESSING using Centralized Server: DDCM should have data consolidated on one or more central servers for one or multiple of the following processes which can be done Real time, or in a batch process
i. Storage of the various processed data on the DDCM machine (emphasis is on storage in a relational database, with the fingerprint image stored in such quality or format that makes it easy to electronically search database based on Fingerprint and/or photograph stored in a format that makes it easy to electronically search database based on picture/photograph)
ii. Check for duplicates or previous entries for this individual on the Central server/storage
which is hence a check across several grouped (or ALL) associated DDCM.
iii. Further processing of the Biometrics to match with other servers e.g. Passport,
national identity, SIM card register, voters register, criminal database e. t. c
2. Wireless distributed interface: By virtue of the sensitive nature of biometrics and data capture machines, Wireless means is the most effective way of movement of data between any point that are not physically close. This can be done directly or using batch processing depending on availability of infrastructure. Wireless distributed interface ensures that multiple mid-processes (like batch processing) can be put into it with the overall objective of transfer of data from between to remote systems such as DDCM to server or between multiple Biometric processing servers is achieved. This can be in 1 or multiple forms as described below.
a. Data Movement: Wireless can range from copying plain or encrypted forms of the database from the DDCM,
i. physically through external storage devices or
ii. electronically via wireless technology. (Patent places special emphasis on this Mode)
Wireless technology however offers the most secure means of data transfer from the DDCM to the server or between several servers. This security is in various forms including confirmation of the identity of DDCM machine before accepting, ensuring the Database being transferred is secure and unhampered with during the process.
b. DDCM Data Processing: In addition to the distributed Biometrics, it is even more effective to combine this with real time direct submission of data from the DDCM (while still in the field or remote locations) to the central servers, to ensure that even attempts at multiple registrations at different locations or different machines are captured and identified at the point of registration on any of the DDCM. This Real time consolidation or transmission also limits data lose in cases of crashed or lost Portable Telecommunication devices.
How to 'PROCESS' CAPTURED 'BIOMETRICS':
a. Finger print Biometric process:
i.Validation and/or storage: A fingerprint contain a number of unique physical characteristics called minutiae, which includes certain visible aspects of fingerprints such as ridges, ridge endings and bifurcations (forks in ridges). Minutiae are generally found in the core points of fingerprints, located near the centre of the fingertips. These characteristics are used to distinguish two fingerprints, or to state that they are the same. Even identical twins have different fingerprints. Even the fingers of the same individual have different fingerprints.
(When a fingerprint is applied to - or passed over - the sensor window of the fingerprint reader/scanner, the fingerprint is scanned and a gray-scale image is captured). The DDCM/software should identify the key minutiae points from the captured biometric finger image. These points are then converted/extracted using a process we call EXTRACTION into a unique digital representation, called "template", comparable to a very big password. If the conversion is done successfully and the template is confirmed to conform to the standard categories /classification of a finger print, the finger print is accepted and stored. Else it is rejected and the registrant asked to re-enrol the finger. We strongly advise the storage of both the template and actual image of the fingerprint though at least one of them MUST be stored. The fingerprint template is not only smaller than the finger image, but also faster and more accurate to process when comparing two fingerprints.
ii. Search for duplicate: Any search for duplicates (MATCHING PROCESS) should be with the use of the Extracted Finger print template. We strongly advise each finger template should be searched across the database of ALL finger templates captured for ALL/each person/individual. A Score method is then used whereby, if the finger meets a benchmark score (which is determined by exact software in use), it is termed a duplicate. This would minimize the False Rate of Acceptance, and False Rate of Reject which are critical benchmarks for rating a good biometric fingerprint process.
The finger print duplicate search/Matching should cater for orientation displacement and angle rotation. i.e. if a fingerprint is upside down or at any angle, it should still effectively match it to the original one that was done proper way up.
b. Photograph /Face Biometric process: The Process for verifying the Picture is two folds.
i. Simple Picture: This can be to simply conform that the picture/photograph fields are not empty, and the content is a valid photo or picture image.
ii. Person confirmation: The Process can attempt to at a minimum confirm validity of the Picture as a person by identifying the Eyes and Mouth based on common characteristics of the opened human eye. It can also go further by analysing the distances between the eyes and mouth in proportion to overall picture size to confirm if the picture meets minimum standards. Also, it can go further by measuring overall facial structure, distances between eyes, nose, mouth, and jaw edges. These measurements are retained in a database and used as a comparison when a Photograph/facial duplicate search is performed.
Having fully explained the intended process (patented) we would also, now explain specific products that will still implement the process above, and still beneficial based on our local environment.
How to PROCESS CAPTURED BIOMETRICS:
Finger print Biometric process:
Validation and/or storage: A fingerprint contain a number of unique physical characteristics called minutiae, which includes certain visible aspects of fingerprints such as ridges, ridge endings and bifurcations (forks in ridges). Minutiae are generally found in the core points of fingerprints, located near the centre of the fingertips. These characteristics are used to distinguish two fingerprints, or to state that they are the same. Even identical twins have different fingerprints. Even the fingers of the same individual have different fingerprints.
(When a fingerprint is applied to - or passed over - the sensor window of the fingerprint reader/scanner, the fingerprint is scanned and a gray-scale image is captured). The DDCM/software should identify the key minutiae points from the captured biometric finger image. These points are then converted/extracted using a process we call EXTRACTION into a unique digital representation, called "template", comparable to a very big password. If the conversion is done successfully and the template is confirmed to conform to the standard categories /classification of a finger print, the finger print is accepted and stored. Else it is rejected and the registrant asked to re-enrol the finger. We strongly advise the storage of both the template and actual image of the fingerprint though at least one of them MUST be stored. The fingerprint template is not only smaller than the finger image, but also faster and more accurate to process when comparing two fingerprints.
Search for duplicate: Any search for duplicates (MATCHING PROCESS) should be with the use of the Extracted Finger print template. We strongly advise each finger template should be searched across the database of ALL finger templates captured for ALL/each person/individual using the one-to-many matching process i.e. 1:N. A Score method can then be used used whereby, if the finger meets a benchmark score (which is determined by exact software in use), it is termed a duplicate. This would minimize the False Rate of Acceptance, and False Rate of Reject which are critical benchmarks for rating a good biometric fingerprint process.
The finger print duplicate search/Matching should cater for orientation displacement and angle rotation. i.e. if a fingerprint is upside down or at any angle, it should still effectively match it to the original one that was done proper way up.
Photograph /Face Biometric process: The Process for verifying the Picture is two folds.
1. Simple Picture: This can be to simply conform that the picture/photograph fields are not empty, and the content is a valid photo or picture image.
2. Person confirmation: The Process can attempt to at a minimum confirm validity of the Picture as a person by identifying the Eyes and Mouth based on common characteristics of the opened human eye. It can also go further by analysing the distances between the eyes and mouth in proportion to overall picture size to confirm if the picture meets minimum standards. Also, it can go further by measuring overall facial structure, distances between eyes, nose, mouth, and jaw edges. These measurements are retained in a database and used as a comparison when a Photograph/facial duplicate search is performed.
Having fully explained the intended process (patented) we would also, now explain specific products that will still implement the process above, and still beneficial based on our local environment.
