You are on page 1of 21

FACIAL RECOGNITION SOFTWARE

AN INFRINGEMENT OF PERSONAL RIGHTS and FREEDOMS


John Alan Sutherland AU # 2980775 POLI 580 Dr. J Smith Fall 2011

ABSTRACT The use of Facial Recognition Software has expanded beyond the simple application of verification and identification of known criminals and terrorists. It is being used to correlate diverse sources including a persons biometric data from pictures and video surveillance, personal non-biometric information and what can be described as generalized group traits into profiles of individuals for the purpose of predicting his or her future behavior. The method of collection of this material without the subjects consent, the misuse of group trait generalizations for individual profiles and the continued discriminatory use of racial profiling in selecting subjects for surveillance infringe the basic rights of citizens in a free and democratic society.

INTRODUCTION In the post 9/11 era Facial Recognition Software (FRS) has become an important biometric tool used by military, security and police forces allegedly to fight crime including terrorism (Volpp, 2002). However the uses of FRS and its usefulness to the state extend far beyond it being a simple tool for identification of known criminals or terrorists. On its most public level FRS is an internet based technology allowing a photograph or a video surveillance taken of an individual to be quickly compared to pictures stored in a data bank for the purposes of determining identity (Thornhill). That part of FRS use is seen as being

completely automated with little human input in the comparison. In this application FRS has its highest level of reliability. But that reliability is still subject to weaknesses in the digital camera technology which is central for comparison between pictures. On another level FRS technology is being used for profiling of individuals who are of interest to the state. FRS analyzes photographs or video surveillance of a person to collect biometrical data. This is then correlated with other personal and general non-biometrical data in order to create a pattern of behavior for a person. That pattern can then be used to predict what his or her intentions are for future action. FRS relies upon multiple sources of information to reach a level of meaningful predictability in this application. In order to complete a profile FRS requires a high degree of human input to fill in the gaps of absent information and to make generalizations about the individual which can affect the validity of the process. Profiling has been described as an attempt to deal with the diversity and complexity of reality through categorization (Backhouse A. C., 2008). Once an individual is categorized into a particular group FRS assigns a generalized pattern of behaviors thought to be common to that group to the individual. Individual behavior patterns are thus skewed by group behavior patterns both positively and negatively. Profiling becomes racial profiling when it classifies persons as belonging to a particular race or ethnic group. Generalized so called racial Traits are then used to profile individuals who have been identified as belonging to a particular race or ethnic group based on their physical appearance. This is by far the least reliable application of FRS and one which critics say is another form of racial prejudice and discrimination.

In attempting to extract biometrical data from pictures there are many factors which can cause distortion of images and hamper accurate matching and identification. These include such factors as lighting conditions, changing weather conditions, complex and/or moving background and complex foreground and occlusions (Angelos Yannopoulos, p. 97). All of these may prevent accurate matches and even result in incorrect matches or false positives. Despite these flaws, governments and private security interests have sold FRS to the public based on the argument that its use provides greater protection against potential terrorists (Magnet, 2007). While it speeds up a process that would normally require checking thousands of so called mug shots of criminals and/ or terrorists is it more accurate? But what of concerns as to the use of this technology to collect data on citizens without their consent. Once a picture of a suspicious person has been taken it is available to the state for a wide range of purposes including future monitoring of persons in whom the state has an interest. Most agents of the state deny that the subject has any right of privacy to a photograph taken in a public place (Angwin, 2011). When FRS is used for detecting and interpreting suspicious behavior in order to discover dangerous intentions (Science and Technology: If looks could kill; Surveillance Technology|, 2008), it crosses into an area which must rely upon nonscientific generalizations in order to create a profile of a person. This process prejudices the individual because of his or her supposed race or ethnic group (Angwin, 2011) . Experience has shown that much of what is generalized about individuals is based on assumptions of traits for different races and ethnicities. A person whose profiling pattern is based on these ethnic or racial generalizations loses the equality of rights including the protection of the state guaranteed to all citizens.

Since the decision to photograph or conduct video surveillance is based on the subjective reasoning of the person initiating the process, it follows that if the subject is a suspicious subject primarily because of race this is racial profiling (Muffler, 2006). Human reasoning is based upon the experiences, training and biases of the individual making the decision to photograph. It is clear that for reasons of economic cost and sheer volume a decision has to be made by someone as to who will be photographed. But if that decision is based on a biased view that some groups of individuals are more likely to be criminals or terrorists than others (Nakamura, 2002)this is not only discriminatory but illegal. The weakness in FRS technology is that it can and often does rely upon groups constructed along racial or ethnic lines (Nakamura, 2007). Taking a picture of a Middle Eastern individual at an airport screening center and comparing it to databanks of potential terrorists is based on a generalization that all Arabs are potential threats to western society. This is racial discrimination and prejudice. It is neither scientifically based nor is it based on any non-biometrical data for the individual. Whether racial profiling is ever justifiable in the interests of national security is another issue. So too is whether individual freedoms should be curtailed to allow this invasion of privacy by FRS technology which denies equal rights to these individuals. The general public seems to have accepted FRS technology without questioning to what extent it infringes upon individual rights and freedoms. This is largely due to the partnership of state and media in raising the level of the terrorism threat to panic proportions. FRSs have allowed individual privacy to be undermined to theoretically benefit collective security (Nakamura, 2008). As warned by Whitaker we are fast approaching the state of total surveillance of our lives (Whitaker, 1999) with it seems very little protest.

This paper examines what FRS does and does not do; and how and why the use of FRS to create profiles and interpret individual biometrics and their biometric patterns can become racial profiling encompassing racial discrimination and lead to increased racism (Fox, 2010-2011). Further it discusses what, if anything, can or should be done to safeguard the equal rights of all citizens from such interference. We are a society that believes in the rule of law. That rule encompasses the principle that anything that creates discrimination against certain groups based on ethnicity undermines a free and democratic society. The diminution of the rights of any group within our society ultimately affects the rights of all. Understanding the basis of FRS technology is important in order to expose its weaknesses and how its use unfairly harms the reputations and rights of ethnic groups of our citizens (Hosein). THE PARAMETERS OF FRS To understand how FRS technology can be abused and lead to unwarranted invasion of privacy and racial profiling it is important to understand what FRS is and what it is not. Initially it was used by military, police and security forces for determining identification of persons who were already known threats to the state. Cynically one can question whether individuals who have their pictures in data banks for past crimes would place themselves in a position to be photographed by security forces. In reality FRS is being used by state agencies for the purpose of gathering further data on persons who may pose a threat to the powers of the state . Digital pictures once taken become part of the states data collection for that individual. In this application FRS technology is used as part of the process to construct, correlate and interpret profiles or patterns of behavior for persons the state wishes to monitor. FRS constructs profiles through use of pictures; enhanced

with additional video surveillance photography; taken over a period of time; supplemented by other personal information and data about the person; and using generalizations about the person based on comparison to other persons with similar biometric traits which we refer to as group profiling. From this composite profile or pattern FRS allows state agencies to discover (Hildebrandt, 2008) what the suspects future actions will be. Will the person be a criminal or terrorist threat to the state? FRS assists in this data mining and by providing patterns of correlations between data in respect of an individual and his or her group. Trust me it is not capable of finding terrorists by capturing the deviant look (Nakamura, 2008). THE PROBLEMS WITH DIGITAL PICTURES It is easy to believe from watching the television show 24 Hours (Nakamura, 2008) to reading the hype in the popular press (Angwin, 2011) that facial recognition devices using digital camera technology are infallible in their ability to quickly match faces to databanks and confirm identity. While a picture is said to be worth a thousand words, a digital camera is not the human eye with a human brain capable of interpreting what the eye has seen. In fact digital is not even the best camera for creating an accurate representation of what the human eye sees. While it is true that faces play an essential role in human communication (Dugelay, 2005) digital photography is probably one of the least effective ways of capturing facial expressions. We are all photographers in an amateur sense. Without any prior training we can take pictures of people around us incidentally recording how the person looked at a particular point in time. The photograph album becomes a family record of our time on earth and that of the people close to us. But many of our pictures are poor images of our subjects. Often when we later bring

out the albums we sometimes have difficulty identifying who the subjects are because of the lack of clarity of the pictures. Digital photographs taken for the purpose of identifying a person or capturing their biometric facial traits are often done in an amateur manner with similar results to our family photographs. Usually the conditions under which the picture is taken do not produce a picture similar to the one stored in a database. The softwares capability for direct comparison between pictures is only as good as the product in the picture: poor picture quality leads to inaccurate comparisons. When pictures are similar but not perfect matches does this not logically lead to subjective interpretation the agents of the state as to whether there is a match? In court proceedings one of the questions routinely asked of a witness presenting photographic evidence is: Is this an accurate representation of the person (or thing)?. To the person who took the picture it is only as accurate as he or she recalls at the time the picture was taken. In the recent riot in Vancouver, British Columbia following the Stanley Cup finals, photographs and videos allegedly taken by amateur onlookers were used to identify individuals who were subsequently charged with crimes against property. Whether these pictures will stand up in court to identify the accused in the act will have to await the trials including cross examination by the defense to highlight any differences between the digital pictures and the actual accused in the witness box. Even if a digital picture is of good quality it is seldom able to display emotion. When we take our own pictures we actively encourage our subjects to show happiness by asking them to smile for the camera. But rarely do our photographs disclose any other emotion. FRS professes not to need professionally taken pictures in order to be effective in identification. But most digital pictures have too much room for error; usually are not good quality ; and fail to be detailed

enough to bring out any facial peculiarities or emotional facial signs. Manufacturers of facial recognition devices oversell the capabilities of their digital systems. Digital pictures taken in the field cannot be compared to those taken by professional photographers. Professional photographers have the skill and expertise to create perfect pictures. They still require proper lighting and time to create them. In fact they are able to create pictures that will portray a person or object in any manner they like and with any emotion they want to display. I once defended an accused in a case of sexual assault where he was identified from a police lineup and picked up the next day on the street near the scene of the attack. The victims description of her attacker was that he had dark features with dark hair, dark clothes and a dark complexion. My client happened to live in the area of the crime. He had dark hair and was a darker ethnic Italian. He fit the victims description . The victims overuse of dark in her description conveyed to me an inability on her part to clearly see her attackers features because of the night darkness in the location of the attack. My professional photographer and I went to the scene at night where he took several pictures of the area. By adjusting the camera lens the area could be made as dark as the professional photographer liked. While the prosecutor objected to the introduction of our pictures as evidence as they were not taken on the night of the attack the judge allowed them in as evidence for identification purposes as an accurate representation of the site of the attack. These pictures showing a very dark area as well as my young clients alibi of being home with his wonderful Italian mother resulted in his acquittal . Clearly photographs have the ability to distort what the human eye can see and produce a representation unlike what was seen. Of course even eyewitness descriptions may lack validity and cause the creation of misleading profiles. There is solid evidence to cast doubt on the reliability of witness

descriptions. In recent years there has been an increasing trend in racial hoaxes, or cases where crimes, real or fabricated are falsely attributed to minority suspects (Gumbhir, 2007). Racial hoaxes prey on larger cultural images and beliefs associated with race/ethnicity and crime, drawing on the publics fears and misgivings to conceal aspects of the overall situation. An example is the widely publicized case of Susan Smith, a white mother of two children who in 1994 accused a black male of carjacking and kidnapping her children when in fact she had murdered them. If the content of crime reporting and witness statements can be influenced by racial/ethnic biases, for whatever reason, then this information should not be embraced without greater scrutiny in deciding who should be the subject of photographic monitoring. Professional photographers are experienced in taking photographs of people with an eye for bringing out certain prominent features of their subject to portray traits which give the person a certain uniqueness. The portrait photographs taken by the famous Canadian photographer Karsh are important examples of that work. His famous portrait of Sir Winston Churchill was taken in a way to convey the strength and determination of the British war-time leader which was one of the strengths of Churchills personality. According to Karsh this was his intention in setting up the picture. Few if any of those manning security cameras have the ability or skill to take such pictures. Therefore pictures as indicated are only as good as the skill of the photographer given time and conditions to create a professional photograph. BIOMETRIC TRAITS AND RACE All of which brings me to the point of whether it is even possible (much less desirable) to interpret race and ethnicity from photographs. Having failed badly in my own attempts at correctly identifying the ethnicities of various Chinese, Japanese and Korean subjects from their

10

pictures at http://www.alllooksame.com I can support those that argue that there is no way that photographs can accurately convey a persons race. And that particular website uses only pictures of non-mixed nationalities. Imagine the difficulty of trying to determine the ethnic makeup from a picture of someone who is of a racially mixed background. Photographs convey nothing more than a mere representation of images frozen in time from which the viewer must draw conclusions. Further the necessity of drawing those conclusions leads the viewer to consult his or her own generalizations or those that have been preprogrammed in a person from the opinions of others. Even more specific and perhaps more accurate than photographs in identification are the use of iris scanners to identify individuals by the biometrics of their eyes. But photographs in and of themselves cannot convey race or ethnicity. Can photographs convey enough in the way of biometric data to be of probative value? The scientific basis for using physical characteristics or biometrics are not new. But what are biometrics? Biometrics have long been used to verify and identify individuals and groups by using what can be described as measureable human characteristics both physiological and behavioral (Angelos Yannopoulos, 2008). Biometric traits are defined as human characteristics that must satisfy certain requirements in order to be useful for identification. These requirements are: sufficient inter-person variability for distinctiveness purposes; invariability over a period of time and thus a rather low intra-person variability; measurability; and universality (A.K. Jain, 2004). Physiological characteristics are strongly connected to the human body. They include such traits as iris patterns, face image, odor, hand and finger geometry, DNA and fingerprints. We often refer to these as hard biometrics. But we are still limited in the number of physiological traits we can use for identification of particular individuals as well by the cost of compiling large databases of individual fingerprints, iris prints and other hard biometrics.

11

What has become more important in the field of biometrics are what are referred to as behavioral biometrics (Angelos Yannopoulos, 2008). These can be described as measurements of human characteristics related to a persons conscious or unconscious behavior, actions, or skills-and not his/her physical features. They are functions performed by a person at a specific point of time for a specific reason (e.g. a persons typing patterns, signature or method of moving the computer mouse) (Vassiliki Andronikou, 2008). Why are these important? Because in order to understand how a person will act in certain situations we need to observe that person and make certain conclusions as to why he or she is acting in a certain way. FRS becomes a process of interpreting a persons actions and trying to discern if there is a pattern to their actions which can tell us what the persons intentions might be. Behavioral biometrics can provide useful profiling information such as a measure of a persons preferences or mood (Angelos Yannopoulos, 2008). More importantly these can link the individual to a nonbiometric profile. These capabilities can be regarded as direct and indirect profiling and constitute behavioral biometric profiling.

PROFILING What do we mean by profiling? Profiling is a process of discovering correlations between data in databases that can be used to identify and represent a subject and/or the application of profiles (sets of correlated data) to individuals and represent a subject or to identify a subject as a member of a group or category (Wim Scheurs, 2008). In simple terms it is a description of who the person is including his or her physiological traits combined with observations of the persons actions as well as generalizations we make from having observed similar people with similar

12

types of traits and actions. A good example is the profiling done on television shows like C.I.S. when the police forensic unit attempts to come up with a profile of the suspected killer based on his or her physical evidence found at the scene such as fingerprints or shoe prints, plus perhaps a witness statement and evidence collected from the wounds to the victim plus similarities to other similar crimes. The Latin term modus operandi also is sometimes used instead of profile. However when we try to classify the person as belonging to a particular race or ethnic group and then ascribe to that person what we believe are general traits of that racial group we are engaged in racial profiling. Does it make any sense to describe an individual in a way that we have designated as appropriate to the traits of that group even though he or she may only partially conform to those traits (Backhouse M. , 2005)? If we believe that First Nation people are lazy we may incorrectly give a person this trait because we believe they are First Nations. Have we not just reinforced our racial prejudice towards that ethnic group? The more complex a system of attributing traits to a group becomes the more prejudices it may end up incorporating into the profile of the individual (Johnson, 2004). Further these are not scientifically proven traits but anecdotal and part of our particular culture. The ascribing of these unscientific traits to a person only serves to reinforce racial discrimination against that person. The use of stereotypes, loose generalizations or broad statistical correlations to suspect a person of being a criminal or a terrorist is racial profiling (Merrick J. Bobb, 2006). The more general the criteria used to isolate individuals, the greater the likelihood that people will be targeted repeatedly or erroneously (Merrick J. Bobb, p. 32). At the core of all racial profiling definitions is the idea of racial discrimination (Gumbhir, 2007).

13

WHY AGENTS OF THE STATE CONTINUE TO USE RACIAL PROFILING Security and police continue to use racial profiling because they can get away with it. They are given a great degree of leeway by most courts in using race as one factor in their profiling of suspects as long as they can show that there is a basis to suspect that a crime has been committed. More objectionable is the concept of race based policing or racially biased policing where the prime reason for photographing or surveillance of someone is because of their race. In general terms every definition of racial profiling includes the idea of racial discrimination or using race as a criterion for treating one person or group differently from others (Gumbhir, p. 16). Racial profiling is the intrusion of racial/ethnic biases into the decision to photograph or conduct surveillance and the improper use of information on race/ethnicity in criminal/terrorist profiling procedures (Gumbhir, p. 16). The reality is that that there are racial/ethnic disparities present in law enforcement practices. The source of racial profiling occurs primarily in officer decision making whereas the institutional aspects of criminal profiling often merely condone such practice. How do we ensure that state institutions monitor and address the misuse of race and ethnicity in individual officer decision-making. In the United States racial profiling is defined as any police-initiated action that relies on race, ethnicity or national origin rather than the behavior of an individual or information that leads the police to a particular individual who has been identified as being, or having been, engaged in criminal activity (Gumbhir, p. 17). This definition is grounded in the theory that it is the use of racial/ethnic stereotypes in officer decision-making that causes differential treatment among members of different ethnic groups. Stereotyping may be related to the overt racial prejudices of individual officers.

14

How is it possible to eliminate racial profiling from acceptable law enforcement tactics which place prime reliance on a suspects behavior and known information on crimes and criminal enterprises? Criminal profiling instructs officers to consider multiple indicators of suspiciousness. It is likely that officers who consider race/ethnicity in their decision making also consider aspects of a suspects behavior as well as descriptions of suspects-at-large in this process. Is the presence of some behavioral indicators or information enough to dismiss claims of racial profiling? The United States Supreme Court has stated that as long as officers can provide some legal or valid reason for their action, questions about stereotyping and judgments based on race/ethnicity are rendered moot. But will the use of ethnic/race stereotypes not cancel out the legal reason for the officers actions (Gumbhir, p. 18)? It appears that most definitions of racial profiling find it acceptable for officers to consider race/ethnicity when they are working from a description of a suspect in a known crime. This result reveals a distinction between reactive and predictive or proactive police action. In terms of reactive police action, working from descriptions provided by witnesses to crimes is a longstanding investigative tool. If the race or ethnicity of the suspect is reported by the victim or witness it logically brings race into the profiling. On the other hand the absence of a suspects description introduces predictive criminal profiling which relies on demographic and behavioral characteristics as indicators of crimes not known to police and opens the door to discriminatory racial profiling. Some critics define racial profiling as hard profiling and criminal profiling as soft profiling. Hard profiling uses race as the only factor in assessing criminal suspiciousness. Soft profiling uses race as only one factor among others in gauging criminal suspiciousness (Gumbhir, p. 20). Is soft profiling a legitimate and acceptable law enforcement practice? It still rests on the idea that racial/ethnic minorities are more likely to commit certain crimes than

15

whites. An example is in the area of drug crimes. Statistics there show the highest percentage of all drug arrests involve black subjects compared to that groups percentage of the total population. However there are flaws in these statistics as they fail to take into consideration social, political, historical, economical, logistical and legal factors. It is wrong to assume that enforcement related data reflects an unbiased estimate of offending. Data on arrests are nowhere close to being a representative sample of offending in a given jurisdiction. The seriousness of the crime, the strength of available evidence, the relationship between the victim and suspect and the suspects attitude towards the police officer all have an impact on whether or not a suspect is taken into custody. In addition apparent differential offending may be thinly veiled differential enforcement. Finally certain categories or behaviors are radicalized i.e. given racial connections. How do we then balance the use of race/ethnicity in officer decision-making with concerns about racial discrimination? How do we differentiate racial profiling from acceptable decision making processes related to criminal profiling? Racial profiling exists at the conceptual connection between race/ethnicity and suspicion. Therefore to define, theorize and study it we must look at the role of race/ethnicity in the social construction of suspicion and the symbolic suspect (Gumbhir, p. 26). Further an historically and tactically oriented approach must be taken to limiting racial profiling in practice. Attempts must be made in identifying patterns in state agents practices that are indicative of the disproportionate application of pretext suspect attitudes to minority individuals. Ways must be incorporated into police and security organizations so that race/ethnicity is not used as the filter through which other indicators of suspiciousness are interpreted. Race/ ethnicity must never be the tipping point in the individual agent decision-making. As well race/ethnicity should never be used as anything other than a confirmatory factor. Information related to race/ethnicity should be considered only after other

16

criteria have been satisfied. In the absence of other key characteristics of a description such as sex, age, height, and weight, race/ethnicity must not be employed as the justification for police action. The initial decision to suspect that someone is involved in illegal behavior is at the core of the controversy surrounding issues such as racial profiling. It is important to determine how a decision is reached to arrive at the symbolic suspect. Officers act and react based on general indicators of suspicion that are likely to be enmeshed with a number of other symbols and meanings. What is the relationship between suspicion and race/ethnicity? There appear to be four possible dimensions of suspicion: criminal potential; danger; priority; and reward. Generalizations about the criminal potential of racial/ethnic groups will inevitably produce selffulfilling prophecies. When considering the element of danger we have to consider the role of politics in determining and directing police action. Moral entrepreneurship not only creates and modifies laws but is involved in agenda setting and prioritization. Hierarchies of importance, both formal and informal inevitably develop within communities and agencies. Finally officers who are successful in enforcing relevant laws will likely receive formal rewards in the form of commendation and promotion. These in turn encourage more arrests. Finally police culture tends to be isolationist and totalizing in the sense that officers rarely socialize with those who are not knowledgeable of and sympathetic towards police work. As a result they are unlikely to encounter much diversity in their lives and experiences outside of work. Police culture is dominated by philosophy that rejects progressive viewpoints of activists and social scientists, especially those that emphasize racism and social justice in their criticism.

17

CONCLUSION Humans successfully use information gathered from the observation of the behavior of others because the human brain enables us to collect and interpret such data. While computer technology is always improving we have not found a way to give computers the ability of the brain to process intuitively this data. Machine intelligence is not able to interpret biometric and non-biometric behavior using the subtlety of social behavior, of rhetoric, of playfulness, or of mock seriousness. Therefore machines are not able to equal human emotion recognition systems in the area of acquisition, extraction and interpretation. Given these shortcomings any technological system that automates the collection of data on individuals and attempts to build their profiles using group generalizations including racial traits will fail without human subjective input. Given these shortcomings for FRS technology we have to ensure that there are proper checks and balances on the use of automated individual profiles to eliminate the influence of racial profiling from the human decisions based on these profiles. Unfortunately profiling especially group profiling ,which by it very nature allows generalizations, will never be free of errors and will when it involves race and ethnic groups never be free of racial profiling. As well the more complex a system becomes the more prejudices it may end up incorporating. Therefore it is important that as we proceed with more profiling using computer technology that we try to gather as much non-biometric data as possible on an individual in order to limit the generalizations that we have attached to him or her in the form of unproven group traits. Such a step would then require that a decision making classification or interpretation be based on this available incomplete non-biometric data. At this point biometric sensors could be used to offer additional information. Biometrics therefore serving as a link to an individuals profile would offer an opportunity to create a trace of an

18

individuals actions , daily activities and transactions. Biometric profiling should be used to fill in the gap not as the basis for building the profile. We should ever be mindful that the storing of captured raw data in databases poses a threat to the subjects privacy, since potential future processing and extraction of additional information, including age, gender, and ethnicity on this data is possible. Finally we must always be cognizant that profiling enables those in power to enhance their power by making ever more precise decisions that benefit themselves rather than the consumer, individual citizen or employee (Koops, 2008). While society is changing not only through technology but through globalization, commodification, anti-terrorism objectives, and trivialization societys ideas of democracy , the rule of law and fairness are shifting as well. Legal protection must focus more on redressing wrongs rather than continuing to focus solely on preventing them. It has become futile to fight against the invasion of privacy by computerized technology. It is the outcome of data processing-the ultimate use of the data and the resulting knowledge that should be the object of protection rather than solely on the collection or processing itself.

19

Works Cited
A.K. Jain, A. R. (2004, January Vol 14 Vol 1). An Introduction to Biometric Identification. IEEE Transactions on Circuits and Systems for Video Technology, special Issue on Image- and Video based Biometrics. Angelos Yannopoulos, V. A. (2008). Behavioural Biometric Profiling and Ambient Intelligence. In M. H. Gutwirth, Profiling the European Citizen: Cross-Discipliary Perspectives (pp. 89-109). Rotterdam, The Netherlands: Springer Science. Angwin, E. S. (2011, July 13). Device Raises Fear of Facial Profiling. Retrieved October 2, 2011, from The Wall Street Journal Online: http://global.factiva.com.ezproxy.lib.ucalgary.ca/hp/printsavews.aspx?ppstype=Article&p Anonymous. (2008, October 25). Science and Technology: If looks could kill; Surveillance Technology|. Retrieved October 2, 2011, from The Economist Online: http://proquest.umi.com.ezproxy.lib.ucalgary.ca/pqdweb?index=0&sid=1&srchmode=1& Backhouse, A. C. (2008). General Description of the \process of Behavioural Profiling. In M. gutwirth, Profiling the European Citizen: Cross-Disciplinary Perspectives (p. 48). Rotterdam, The Netherlands: Springer Science. Backhouse, M. (2005). Descriptive Analysis and Inventory of Profiling Practices, FDIS Deliverable 7.2. Brussels: European Union IST FIDIS Project www.fidis.net. Dugelay, A. C.-L. (2005). Techniques for Face Motion and Expression Analysis on Monocular Images. In N. S. Strintzis, 3D Modeling and Animation: Synthesis and Analysis Techniques for the Human Body (pp. 201-234). Hershey, PA: IRM Press. Fox, D. (2010-2011). The Second Generation of Racial Profiling. American Journal of Criminal Law Vol 38 No 49, 49-79. Gumbhir, V. K. (2007). But is it Racial Profiling? Policing,Pretext Stops, and the Color of Suspicion. New York: LFB Scholarly Publishing LLC. Hildebrandt, M. (2008). Defing Profiling: A New Type of Knowledge? In M. Gutwirth, Profiling the European Citizen: Cross-Disciplinary Perspectives (pp. 18-26). Rotterdam, The Netherlands: Springer Science. Hosein, G. (n.d.). Threatening The Open Society:Comparing Anti-Terror Policies and Strategies in the US and Europe . Privacy International. Johnson, M. (2004). Biometrics and the Threat to Civil Liberties. Computer Vol 37 No 4, pp. 90-91. Koops, B. (2008). Profing and the Identity of the European Citizen. In M. Gutwirth, Profiling the European Citizen: Cross-Disciplinary Perspectives (p. Chapter 15). Rotterdam, The Netherland: Springer Science.

20 Magnet, K. (2007). Communication Research and the Study of Surveillance. communication Review Vol 10, 277-293. Merrick J. Bobb, N. H. (2006). Racial Profiling. In S. J. Muffler, Racial Profiling: Issues, Data and Analysis (pp. 31-39). New York: Nova Science Publishers Inc. Muffler, S. J. (2006). Racial Profiling: Issues, Data and Analysis. New York: Nova Sciences Publishers, Inc. Nakamura, L. (2002). Cybertypes:Race, Ethnicity and identity on the Internet. New York: routledge. Nakamura, L. (2007). Digitizing Race: Visual cultures of the Internet. Minneapolis: University of Minnesota Press. Nakamura, L. (2008). Interface of Identity: Oriental Traitors and Telematic Profiling in 24. Camera Obscura, 109-133. Thornhill, T. (n.d.). The sinister New software that can scan Any Face and ID it from the Web in Minutes. Retrieved September 30, 2011, from Daily Mail Online: http://www.dailymail.co.uk/scientech/article-2043786/Snap-unhappy-Facial-recognition Vassiliki Andronikou, A. Y. (2008). Biometric Profiling: Opportunities and Risks. In M. H. Gutwirth, Profiling the European Citizen: Cross-Disciplinary Perspectives (pp. 131-145). The Netherlands: Springer Science. Volpp, L. (2002). The Citizen and the Terrorist. UCLA Law Review Vol 49, 1576. Whitaker, R. (1999). The End of Privacy: How Total Surveillance is Becoming a Reality. New York: The New Press. Wim Scheurs, M. H. (2008). Cogitas, Ergo Sum,The Role of Data Protection Law and Non-discrimination Law in Group Profiling in the Private Sector. In M. Gutwirth, Profiling the European Citizen: Cross-Disciplinary Perspectives (p. Chapter 13). Rotterdam, The Netherlands: Springer Science.

You might also like