Sam Mil'shtein

Fingerprint Recognition Algorithms for Partial and Full Fingerprints

An urgent need to develop accurate biometric recognition system is expressed by governmental agencies at the local, state, and federal levels, as well as by private commercial companies. Fingerprinting is the most practical and widely used biometric technique. The pattern of ridges and valleys of each fingerprint is unique. The minutiae based algorithm is widely used for fingerprint authentication. One of the significant parts of this algorithm is the classification of fingerprints which allows minimizing significantly the number of fingerprints referenced for each identification procedure. However, the minutiae algorithm has some serious drawbacks. If the core of a fingerprint is not visible, then identification cannot proceed. Yet in some cases, partial fingerprints need to be identified. We recently developed a novel contactless line scanner for recognition of fingerprint pattern that converts a three dimensional object like a finger into a two dimensional image with minimal distortion. This novel imaging technique based on a line by line scanned image required the development of a new recognition algorithm. In this study, we propose two new algorithms. The first algorithm, called the spaced frequency transformation algorithm (SFTA), is based on taking the fast Fourier transform of the images. The second algorithm, called the line scan algorithm (LSA), was developed to compare partial fingerprints and reduce the time taken to compare full fingerprints. A combination of SFTA and LSA provides a very efficient recognition technique.

Line Scanner for Biometric Applications

Fingerprints are the most common biometric measure taken, and in recent years electronic fingerprint scanning has become commonplace. Although contactless fingerprinting methods are used, it is more common that the fingerprint is taken by pressing the finger against a computer scanner. In the current study, the advantages of the contactless technique are combined with the very promising attributes of line scanning. Line scanner views each portion of the finger perpendicularly, therefore removing the projection errors of conventional flat scanning. The three dimensional image of a finger is recorded in pixel-thick lines by scanning the camera around the finger. The final image captures an uncoiled view of the finger. Emerging technology for biometric identification is based on storing the blood vessel map in the finger. We developed a system where line scanning of finger ridges in conventional light is taken as the camera goes clockwise around a finger and in the Infrared light the image of blood vessels is recorded as the camera goes back to its initial position. Line scanning also opens a new possibility in face recognition where the comparison of major anthropometric lines would help to immunize the recognition system against benign cosmetics as well as intentional efforts to defeat such systems through the use of make-up or plastic surgery.

Mobile system for fingerprinting and mapping of blood - Vessels across a finger

We present the design of a novel, mobile, and contact-less, fingerprint line scanner and blood vessel mapper. This system provides a means for acquiring fingerprints with less than 1% distortion in extremely high resolutions, presenting an ideal technology for high-security and individual identification applications. We envision a difficult scenario where patrolling law-enforcement officers would use this miniature fingerprinting system. In this scenario, a police officer will stop a driver for a traffic violation and require the driver to put one (or a few) of his fingers on the testing equipment. The drivers fingerprint will be stored by the mobile system and instantly transferred through a wireless connection to a laptop located in the cruiser. Simultaneously, the laptop will wirelessly send the collected information to a data center. It is expected, that before the officer completes his written report he will receive on the laptop the information about the drivers criminal record (if any). The mobile fingerprinting system is compatible with AFIS and APIS system. The current study describes advantages of line scanning and challenges of the mobile, miniature design.

Can contactless fingerprints be compared to existing database

The major development of fingerprint technology in the form of wet-ink fingerprinting was initiated and improved for forensic applications by Scotland Yard about 100 years ago. In recent years, new fingerprinting methods such as live scan and contactless methods were developed and continue to evolve. Now-a-days, compatibility between three fingerprinting libraries, namely wet-ink, live-scan and contactless becomes to be a critical problem. The large numbers of people travelling and crossing the borders are subjected to fingerprinting. The database of travelers is much larger than criminal one. In this environment, the need to produce speedy comparison results starts to be problematic because it is done by human operator to provide reliable judgment. From the technical standpoint, fingerprinting methods have improved rapidly to an extent where wet-ink fingerprinting has now been replaced by computer scanner generated images of finger, i.e. digital imaging, and recently developed contactless line scan. In current study, we developed certain steps which will make contactless fingerprints comparable to fingerprints in the other two libraries. The proposed steps are applicable to any contactless fingerprinting technique. The rate of success in comparing these three libraries is 82%. The somewhat low accuracy is introduced by the fact that in the contactless methods the position of the finger is free and not necessarily repeatable. To the contrary, contact based wet-ink and digital imaging do handle a finger in the same repeatable position. Therefore, improving designs of holders for examined fingers in contactless setting should improve the rate of success.

Wireless communication supported by contactless fingerprints

The biometric authentication technology is booming in a variety of applications such as security of computers, cell phones, credit cards etc. The most widely used method of biometric identification is dactylography, as the large database already available and law enforcement agencies have long-term experience with this technology. Widely used methods of prerecording the fingerprint could be as simple as pressing a finger first against wet-ink-pad or dry-powder-pad and then against a white paper or as sophisticated as pressing a finger against optoelectronic reader. In all cases variation of pressure risks unrepeatable deformation of a fingerprint and in turn difficulties in identification of the same very finger. In current presentation we discuss design of the software which allows cell telephone in sequence to read, to binarise, to digitize and to recognize a fingerprint without modifying basic hardware of a telephone. Depending on the design of the telephone one or both cameras might be capable to pick up the detailed image of the fingerprint when the finger is at the very close distance, about 2-3 centimeters, to the camera. Moreover the developed software allows picking up the high resolution image of the fingerprint. The cell phone generates a sound click when the fingerprint image is taken. In about 20ms the fingerprint image and its binarization are erased leaving the digitized equivalent of the binarised fingerprint in the memory of a telephone and memory of a server located at private company or government site. The reconstruction of an image of real finger presents a very challenging task making ability of hackers interference into such communication almost impossible.

Biometrics assisted secure network transactions

Biometric methods of authentication such as fingerprints, hand geometry, face recognition, voice, retina and iris scan are being increasingly used for security purposes. However online transactions still use traditional knowledge based, i.e., Personal Identification Number (PIN) or password for authentication. Such methods have proved ineffective over the course of time. Traditional key based authentication has been in use over the years, but not many forays have been made in to securing network transactions using biometrics. The length of the authentication key used in authentication systems is an important piece of the authentication process. In non-biometric authentication processes such as passwords or PIN numbers, depending on the length of the key, the information may be vulnerable to access by unauthorized users. Our system makes conventional key based authentication obsolete by replacing or supplementing the normal key characters with a personal identifier of the user. Without this biometric key the service is inaccessible. In this study we propose novel authentication protocol, based on use of a contactless undistorted fingerprint images to generate authentication key. The authentication process will be reinforced by randomly selected areas of a fingerprint that are used to generate keys. An additional layer of security is provided by using blood vessel map to determine a live finger is being used. We assert that usage of contactless, thus non-distorted, fingerprint as input to generate single session keys will enhance security of network transactions because of the amount of information in randomly selected segments of a fingerprint which is available to generate unique keys.

Development of student innovation skills through industry-university research collaboration

When university groups are involved in joint R&D projects with industrial companies, students learning is intensified and focused on practical aspects of design and fabrication of electronic components and systems. In current presentation we discuss the development of specific skills which undergraduate and graduate students of UMass do acquire in development of RF technology working closely with engineering teams of companies such as MACOM Technology Solutions, RF Micro Devices, Raytheon and others. Students coming to a company site should possess a strong knowledge in solid-state and quantum electronics, RF circuitry design, and device processing. As interns with a company their learning starts with discussion of potential products under development. Before sending a wafer to processing students need to understand the entire set of technical parameters, and processing technology used in production. After first experimental wafer is tested, the most crucial step is to find the ways to improve the performance of a product by re-designing the component and the circuitry, attempting to change the fabrication, if possible. It is at that step the creative thinking most often generates new ideas and patents.

Perspectives and limitations of touchless fingerprints

In recent years, fingerprinting technology has developed at a rapid pace. Fingerprinting is still the most prevalent method of verifying an individuals identity because every person, has unique fingerprints. However, fingerprinting technology does have its limitations. The perception that fingerprint, as biometric trait, does not change with time is not correct. Aging and medical conditions of an individual do impact the overall shape of a fingerprint. Current study reviews the state of the art in fingerprinting technology emphasizing limitations of fingerprints produced by contact methods versus contactless techniques, describing drawbacks of the hardware as well as the limitations of software.

Developing Student's Innovation Skills for Globalized Electronic Industry

One of the very dynamic and demanding industries is, with no question, the electronic industry. It is also one of industries strongly shaped by globalization and therefore requires that its engineers to carry multi-faceted training. To be hired and capable to keep the job the today and future engineer must demonstrate innovative skills or, in more general terms, to be resourceful. This 10 years long study presents the results of training of two groups of electrical engineering students selected each academic year to develop design projects. The requirement of each two semester long project is to develop novel circuitry (amplifiers, switches, etc.) or novel component (transistor, photodetector, etc.). Both groups were provided with commercial design packages. The difference however is that students of one group were supported by internships and worked at the sites of electronic companies and the other group was supported by university research assistantships. Every year based on scholastic achievements 10-15 students were selected for the project oriented course. Students were given the freedom to select a project topic. In current presentation we discuss the development of specific skills which undergraduate and graduate students of UMass do acquire in development of RF technology working closely with engineering teams of companies such as MACOM Technology Solutions, RF Microdevices, Raytheon and others.

Photodetector with uniform response in 620 nm to 870 nm range

In this study, p-n junction, which consists of a p-type region of In_0.4Ga_0.6P 5 mum thick layer doped to the level of N_A= 10¹⁴ cm^-3 is developed. The n-type region is doped to the level of N_D= 10¹² cm^-3 and consists of modulation doped layers ranging from In_0.4Ga_0.6P to In_0.9Ga_0.1P. And photodetector with an approximately uniform response in the near-infrared spectrum between 620 nm to 870 nm is designed.