The Use of
Technology Assessment
Technology Assessment Report prepared by:
Table of Contents
II. Iris Recognition
III. Retinal Recognition
IV. Hand Recognition
VII. Speaker Recognition
VIII. Policy Options
X. Conclusion
XI. Works Cited
When looking into different practices of biometrics one must fist become familiar with what defines biometrics. Nita-Rotaru defines biometrics as “automated methods of recognizing a person based on a physiological or behavioral characteristic” (p.4). These physiological and behavioral characteristics can range from voice recognition, hand and finger recognition, and iris and retinal recognition.
The iris can be described as the middle part or coat of the eye called the uveal which has a diaphragm that is stretched across the outside of the eye (Wildes, 1997, p.1348). What makes this part of the eye a valuable tool in visual recognition? Wildes writes that “the special patterns that are apparent in the human iris are highly distinctive to an individual.” In fact, Microsoft Encarta Encyclopedia states that “each iris is unique, and even irises of identical twins are different” (Microsoft® Encarta® Encyclopedia (2003). Being able to distinguish between all individuals is very important to a recognition system, but happens when a person tries to alter their appearance? Clinical observations have shown evidence that after early childhood growth the iris maintains a distinctive pattern that is highly recognizable (Wildes, 1997, p1349). In order for the recognition system to work there must be data collection, or image acquisition, that must take place.
In the case of the iris recognition Wildes states that there are several important factors which include: first, “sufficient resolution and sharpness to support recognition. Second, it is important to have good contrast in the interior iris pattern without resorting to a level of illumination that annoys the operator, i.e., adequate intensity of source constrained by operator comfort with brightness. Third, these images must be well framed without unduly constraining the operator (p. 1351).” All of the acquisition must be taken with minimal inconvenience to the operator as well as the test subject. All of the equipment must be precisely set at the proper distance for acquisition as well as minimizing all the background reflections that could disturb the image making it unreliable for recognition (Wildes 1997 p. 1353)
The retina, see fig. 2 right, can be described as a layer of complex blood vessels and nerve cells (Microsoft® Encarta® Encyclopedia (2003). The complexity of this layer of the eye makes this form of biometrics one of the most reliable forms of verification (Nita-Rotaru, 2003, p.22). The key to retinal recognition involves low intensity lights, see fig. 3 below, shined directly at the test subject’s retina, thus making the precise location of the subject along with the perceived public opinion on safety the disadvantages of retinal recognition (Nita-Rotaru, 2003, p.22). Another disadvantage that has lead to retinal recognition not being implemented into a wide variety of applications is the expense that installing and maintaining this system entails (Nita-Rotaru, 2003, p.22).
Fig. 2
Fig
3
http://www.eyesearch.com/normal.retina.jpg
http://static.howstuffworks.com/gif/identity-theft-retinal-scan.gif
Handprint Recognition scans the outline or the shape of a shadow, not the handprint. It can is used for many types of access, but accuracy does tend to be a problem. Since it is a fast and semi-reliable method of verifying identity many companies use it, but many people have similar hand shapes and sizes so this system is not considered 100% secure. The National Center for State Courts States the use of hand geometry is not enough to identify an individual, but by combining various individual features like fingerprints, this method can be very effective. (NCSC, 2002).
fig 4
Hand Geometry is good for places where quick verification is a positive. (http://biometrics.cse.msu.edu/hand_geometry.html) Scanning a hand takes anywhere from 5 to six seconds. As stated before although it is not a completely secure method of identification it is very fast, and combined with other areas can be very affective.
Fingerprints is an area in which there have been many new and exciting developments in the past several years. Fingerprints constitute one of the most important categories of physical evidence and are among the few that can be truly individual. (Henry C. Lee, Advances in Fingerprint Technology)
fig 5
Fingerprints were first recognized as unique in 1684, recent advances in computing power have given us the ability to capture and compare one fingerprint against another at a user level. Efforts to develop automatic fingerprint identification systems were initiated in the early 1960's in at least three Western countries: The United States, France, and Great Britain. Again the reason for this push was the availability of the digital computer. There was hope that this new technology could assist or even replace the labor-intensive processes of classifying , searching, and matching that are involved in using fingerprints for personal identification. The Federal Bureau of Investigation sponsored research in automatic fingerprint identification in the United States. This began on the initiative of far-sighted Special Agent Carl S Voelker. (
fig 6
Fingerprint recognition compares a user’s fingerprint to a previously stored template and determines validity or authenticity based on this comparison. The template is created from tiny points called minutiae—based on the position of end points and junctions of print ridges—extracted from the fingerprint during enrollment. The comparison of attributes are carried out using complex algorithms during verification.
Facial scan biometrics is an automated way of identifying a person by their distinct individual facial features. Facial scans have recently become a growing concern in this nation. They are being used to find and determine anyone who is known as a possible threat. However this is not their only function they are used to identify and verify people for many different applications. Facial scans are done in many different techniques, and involve advance software to analyze and breakdown specific details and features of each face. Even though the idea of a face scan and the software used to complete them may be complicated they can be achieved with a very simple store bought camera (Biometric ID’s, n.d.). Although there are several different types of facial scans they all use the same basic steps and procedures in a similar way. The steps are:
“Capture – A raw biometric is captured by a sensing device.
Process – The distinguishing characteristics are extracted from the raw biometric sample and converted into a processed biometric identifier record.
Enroll – The processed sample (a mathematical representation of the biometric) is stored or registered, for comparison later during the authentication.
The verification – matching the sample against a record.(Lennon E. 2001).” (Richards, 2003)
Three types of facial recognition techniques used: eigenface, eigenfeature, and thermal imaging (Biometric ID’s, n.d.). Some things sued in these are the three main parts of the face that are consider to not to usually change: “[upper sections of the eye sockets, area surrounding cheek bones, and sides of mouth.]”(Henry, 2001)Eigenface systems capture the image and changes it to light and dark areas. This is also captured in a two-dimensional form in both the initial facial image and the facial image in question. Then the two images are compared according to the points of the two eigenface images (Biometric ID’s, n.d.). Eigenfeature image systems work in the same way except it picks out certain features and calculates the distances between them. The points are the facial features such as eyes, nose, mouth, bone curves, and other distinct features. However, many faces do change over the course of a person’s lifetime so the images in storage need to reflect that. Some images systems to account for this change, but do not always do so correctly. This is where thermal imaging takes over. Thermal imaging takes a thermal image of the face that focuses on the blood vessels because it is believed that if the face changes your blood vessels do not. To accomplish this a infrared camera must be used instead of any type of traditional camera. 3D imaging on both the face in question and the stored face can help to bring out frauds in facial scans.
Figure
7
Some problems with face recognition are keeping information secure to avoid fraud, necessity of cooperation to obtain scan for database, difficulty capturing good image of face at the right angle to compare to database, and ethical issues such as privacy (Richards, 2003). Some advantages are no contact is needed which means a suspect can be identified without their knowledge (Advantages, n.d.). Also ordinary light is sufficient to complete the process and can be used for security purposes and identifying threats.
Speaker Recognition is an automate system that analyses and interoperates a human vocal patterns and characteristics to recognize words, identify, or verify ones identity. This is a more recent biometric and has not been implemented as long as many others (Cabrera, n.d.). This is classified as a behavioral biometric and analyses many different parts of the human speech. Some areas analyzed include language, speech pathologies, and physical and emotional state of the speaker (Graevenitz, n.d.). Speaker recognition technology is one of the cheaper biometrics because it can work with any microphone such as a microphone that comes with or could be plugged into your computer or a telephone handset (Graevenitz, n.d.). However, the better the microphone the higher quality and better the accuracy to the system. Some applications for speaker recognition are recording attendance, granting access to sensitive areas, over the phone verification such as banking, and forensic purposes. Currently the over the phone speaker recognition is the most commonly used today.
There are two different types of information that exist with speaker recognition, low-level and high-level (Graevenitz, n.d.). High-level deals with characteristics that humans can also use to recognize and distinguish for person form the next. Some examples of these are dialect, accents, talking style, and content. Low-level information is used mainly in speaker recognition system to analyze speech. These are pitch periods, rhythm, tone, spectral magnitude, frequencies, and bandwidths.
Speaker recognition systems are classified in many different ways, one distinction that is used to classify these is text dependency. There are two different types text-dependent and text independent (Graevenitz, n.d.). Text-dependent or fixed text systems require to speaker to say a certain phrase or password. This text usually comes from a training system that uses words to better clarify the person’s individual characteristics. To accomplish this a person must first have their voice recorded repeating the chosen word or words. Text-independent or free text has no required word or phrase that must be said. Any speech can be captured and analyzed. This has several advantages over the text- dependent system. First there are no passwords to remember so a person does no have to struggle with remembering one. Second, this helps to eliminate possible fraud because the phrase is not the same every time and cannot be recorded and perfected to grant fraudulent access. However, this text-independent has many problems with mismatches and granting authorized people access. This is why it still mainly remains in the experimental stage.
Another way to classify speaker recognition is speaker identification and speaker verification systems (Rodman, n.d.). Speaker identification is a system that attempting to identify a speaker from a file of known speaker characteristics to determine who if a known person is speaking. This can be done from a person speaking alone or in a group environment. To accomplish this the speaker voice must have been previously recorded and analyzed. Speaker verification is a system that is trying to confirm if a person is who they say they are. This is done by a person stating a password or their identity into the microphone (Technical Information, n.d.). False rejection and false approval are two types of error that can occur when using speaker verification. False rejection is not allowing an authorized person access to the area. False approval is allowing access to a person who does not have access to the area. Both of these are disadvantages that can cause problems in the application it is being used in.
There are several things that can effect and cause problems in a speaker recognition systems (Graevenitz, n.d.). The environment can be an issue for concern when using a speaker recognition system. Background noises in the environment cause the system to result in failure. Error in the speaker’s pronunciation of the phrase or password can cause problems. A person’s mental and physical health variations can cause problem in recognizes their voice patterns on a consistent basis. The speed and attitude can also have affects on the process. All of these can cause problems if the speaker in not in the same environment, condition, and mood as when they recorded their voice.
Figure 7: Flow Chart of a Speaker Recognition System. (5)
Policy Option 1: Distance Education
The National Center for Education Statistics, see fig. 4 below, notes that over three millions students across the nation are currently enrolled in distance education (Tabs, 2003, p.5). With this many students taking advantage of the new technology at hand, critical questions must be raised to assure the validity of the learning taking place by the students. How is it possible for school administration to verify that the students who have
signed up for the distance education classes are actually performing the work and not allowing others to complete the work? This is how biometrics comes into play.
A proposed policy to bring validity to distance education would be to mandate the immediate installation of biometric verification hardware on all computers that are being used to compose and transmit information from the student to professor. By using this hardware and software package it would help to reduce the amount of academic fraud taking place in distance education. Also the increased amount of validity will help the academic institutions to maintain accreditation with its member institutions.
Beginning in the Fall semester of 2005, all students signing up for a distance education class will be required to purchase, or rent from the educational facility, the hardware and software necessary to validate the user of the computer. The otherwise unspecified hardware software combinations will be selected by the particular learning institutions. The cost to the student for the equipment is not to exceed $200.00 dollars per unit. If the institution decides on selecting a hardware software combination that exceeds this amount, the remainder of the expense will be up to the institution to fulfill. The additional charge to the student will be listed as a technology fee.
In the interest of perceived public safety and ethical issues, only fingerprint verification will be used for data collection. Research has shown that this is the least invasive procedure for recognition and verification. Furthermore, any institution choosing not to comply with the recommendation of this legislation shall loose federal funding as well as risk accreditation loss from it member institution.
fig 9
Policy Option 2: Ethical and Privacy Issues
When implementing a new policy into and educational environment the plan must be very thorough and through because of the sensitivity of environment. Depending on the level educational the community, parents, and students are very decisive about would should and should not be done in schools. With the issue of biometrics this decisiveness and sensitivity exists at a very powerful level involving ethical and pricy issues. A policy to introduce biometrics into an educational system to create the least ethical and privacy issues would be to implement a finger print identification program into the school systems daily routines. Some places to implement the fingerprint identification would be lunch payment methods, attendance logs, and entering school sponsored transportation. Implementing finger print scans into to these areas with care control programs would raise the safety and efficiency of the school system. Finger print scans are more efficient than ,any other types of biometrics. As stated earlier voice scans can be affected greatly by the environmental background noises. Schools have many students and many activities going on controlling the background environment can prove to be a incredibly tough task. Face recognition as explained earlier can have problems orientation of the picture and have trouble with growth changes. Also are not as quick as fingerprint identification. Any type of eye scan can be a very efficient way of identifying, but can bring up ethical and safety issues. Currently there are no known harms or side effects of eye scan. However this may not be the whole publics view on the situation. Also the future could reveal a harm that is not known yet. Left open possibilities like these are an extreme cause for concern with parents and society. There are many schools in the nation that already use finger print identification on either their students or staff. Currently these programs are voluntary so not to cause any unwanted controversy by parents whop do not want their kids participating. There are currently 35 schools in the state of Pennsylvania that use fingerprint identification in their school cafeterias (Weekly. 2001). The cost is relatively an expensive costing the a school with a Food Service Solutions program about 4,000 to 5,000 dollars a lunch lane with the scanners and software costing about 900 dollars each. So the costs depends on the schools capacity and own digression in the cost. This expense can be split on a 50/50 basis with the government for any school that decides to join in on the program. This will make implementation a more realistic goal even for school with a low budget. According to Barry Steinhardt, director of the American Civil Liberties “The same authentication can be done with much lower-tech models, and without risking making a database for other purposes.” (Weekly. 2001) This state is not well thought through having lower technologies such as a meal card is more susceptible to loss and fraud. If a card is lost a student can use it fraudulently before it is reported lost. Also have an employee can check the card can cause errors by lack of checking properly where as a fingerprint cannot as easily be faked. Also this method protects the identity of those who receive free funded lunch. Tawanda Worthy a 13 year old student prefers the finger print system because “You don’t have to bring lunch mo0ney, so somebody can’t take it,” (Weekly. 2001). However many questions arise when this type of information is contained digitally such as who will have access, will the system be linked to the internet, and who will monitor the people with access. A Michigan attorney ruled against biometrics because she argued it violates the state’s Child Identification and Protection Act of 1985. This is due to many of the unanswered questions stated above. A school district in Philadelphia has implement finger print identification with their school employees to track when they are there, considering continuing this including their teachers (Borja. 2002). Many of the questions raised above have also been raised in this situation. Ted Kirsch the president of the Philadelphia Federation of Teacher does not trust the school and feels it violates the privacy of the employees. Some other areas of the school system that would benefit form finger print identification are attendance and transportation. Using the fingerprint identification as the student enters the classroom would give a record of where they are and limit a teacher’s mistakes concerning whether student was there or not. Having identification before enter school transportation would record the student entering the bus and acknowledge that they are an enrolled student. This would increase the safety of the students. There are many advantages to implementing a plan that would use fingerprint identification in school system, however alternate plans need to be put in place as well to ensure privacy for the people involved. Also a program should be put in place to inform the student, the parents, and employees to make them feel comfortable with the new system. When new technologies are put into use there are always questions, concerns, and doubts. The most important part in the success of biometric is addressing ethical and privacy issues to make users comfortable and confident that their personal information will be protected.
Policy 3: Implementing Fingerprint Readers in Primary Schools
A proposed policy in implementing fingerprint scanners in all primary schools is being set forth due to society's need for security and proper tracking of all students. This system has been pilot tested in a few schools throughout the country and the results have been positive. This policy will force all schools to incorporate finger scanners into their daily routine within the year 2008. The role these scanners will play will be vital in the future of schools. Having these scanners will allow the school to document everyone who steps into and out their school, making it easier to resolve unwanted person problems if they would happen to occur. These scanners would also be essential in taking attendance in the classroom. With today's stress on teachers and students to reach certain goals, the time it takes to take attendance could be better used to educate the students. This is why the policy is recommending scanners outside the main entrance of the school and at the entrance of all school rooms. This will allow a school to track the entrance of all students and faculty in all school rooms. This information can become vital when an unwanted problem would occur. Also this policy is recommending using fingerprint scanners during school lunch. With this system all students would have to use the system instead of paying for the school lunch. The students will essentially be using their finger as a virtual debit card. This will eliminate the problems of stolen lunch money, lost lunch money, and the stigmatism students receive who are placed on reduced or free lunch.
fig 10
fig 11
There will be grants and government incentives to implementing this system as soon as possible. Schools will be able to apply for grants and they will be given due to school need and the quality of the written grant. Also the government will provide tax breaks to all schools who have this program implemented before 2006. With the price of scanners reducing dramatically the total cost of 100 scanners (an average of what a school may need) will cost around $300,000. Although this may seem like a great deal of money, a school could only get 12 scanners a year ago for the same amount. With this type of decrease in prices the ability for schools to implement this technology is well within reach, and will be expected in the years to come.
Other areas that are recommended to use this system are libraries, school buses, and extra curricular activities. Libraries can use this system to track who checks out what books. Many schools ask students to carry library cards or a school ID, these can be lost or misplaced making it hard for a students to get learning material. This essentially defeats the purpose of the library and ends up hurting the students education. With school buses these systems can be used to document when a student steps on and off a bus. This will be just one more step in making sure that the students safety comes first. It is also recommended if possible to use these systems before and after extra curricular activities. This will allow the school and the parents to keep track of the students and will help eliminate problems that occur in the school after school hours.
The hope is all schools will be able to perform these task in the next four years. These systems will make schools a safe organized place and will ensure the safest environment for tomorrow's students.
Wildes, R. E.
(1997). Iris Recognition: An Emerging Biometric
Technology. IEEE, 85(9), 1348-1351.
Nita-Rotaru, C.
(2003, Fall). Cryptography CS 555. Department of
Computer Sciences Purdue University, West Lafayette, IN.
Microsoft
Encarta (2003, April). Microsoft Encart Online
Encyclopedia. Retrieved April 17, 2004 from , Web site:
http://encarta.msn.com/
Tabs, E.D.
(2003). Distance Education at Degree-Granting Postsecondary
Institutions 2000-2001. National Center for Education
Statistics, July 2003, iv.
Advantages of Facial Recognition (n.d.).
Retrieved April 25,
2004 from , Web site:
http://www.anti-terrorism.com/products/frs/advantages.htm
Biometric ID's (n.d.). Face Scan. Retrieved April 25, 2004 from
, Web site: http://www.biometricid.org/face.html
Cabrera, Antonio Jr. (n.d.).
Biometrics. Retrieved April 25,
2004 from , Web site:
http://www.employees.org/~junior07/
Graevenitz, Gerik Alexander von (n.d.).
About Speaker Recognition
Technology. Bergdata Biometrics GmbH, Germany
Henry, Veronica (2001). Biometrics Face Recognition T echnology. GIAC.
Jain, Anil, & Kulkarni, Yatin (n.d.).
Speaker Verification.
Retrieved April 25, 2004 from Michigan State University, Biometrics
Web site:
http://biometrics.cse.msu.edu/speaker.html
Richards, Sarah (2003). Face
Recognition . Retrieved April 25,
2004 from , Web site:
http://www.csm.uwe.ac.uk/~jgilliga/UQI133H3/digsig/Face.htm
Rodman, Robert D. (n.d.). Speaker
Recognition of Disguised Voices: A
Program for Research. Retrieved April 25, 2004 from North
Carolina State University at Raleigh, Department of Computer
Science. References
Advantages of Facial Recognition (n.d.).
Retrieved April 25,
2004 from , Web site:
http://www.anti-terrorism.com/products/frs/advantages.htm
Biometric ID's (n.d.). Face Scan. Retrieved April 25, 2004 from
, Web site: http://www.biometricid.org/face.html
Cabrera, Antonio Jr. (n.d.).
Biometrics. Retrieved April 25,
2004 from , Web site:
http://www.employees.org/~junior07/
Graevenitz, Gerik Alexander von (n.d.).
About Speaker Recognition
Technology. Bergdata Biometrics GmbH, Germany
Henry, Veronica (2001). Biometrics Face Recognition T echnology. GIAC.
Jain, Anil, & Kulkarni, Yatin (n.d.).
Speaker Verification.
Retrieved April 25, 2004 from Michigan State University, Biometrics
Web site:
http://biometrics.cse.msu.edu/speaker.html
Richards, Sarah (2003). Face
Recognition . Retrieved April 25,
2004 from , Web site:
http://www.csm.uwe.ac.uk/~jgilliga/UQI133H3/digsig/Face.htm
Rodman, Robert D. (n.d.). Speaker
Recognition of Disguised Voices: A
Program for Research. Retrieved April 25, 2004 from North
Carolina State University at Raleigh, Department of Computer
Science.
Secure Biometric Solutions (n.d.). Face Recognition
Technologies. Retrieved April 25, 2004 from , Web site:
http://www.biometricsolutions.com.au/Face%20Recognition.htm