Difference between revisions of "Eye Tracking"
| Line 62: | Line 62: | ||
<h1> Privacy Ethics </h1> | <h1> Privacy Ethics </h1> | ||
| − | The early stages of eye tracking had very little regard for privacy concerns. Since the data was used primarily to create scientific inferences, the data was seen as an open tool to researchers. The benefits to the researchers in the fields of cognitive science, education, and health often outweighed to privacy concerns associated with access to data sets. However, until this point very few methods have been proposed. <ref name = "privacy">Liu, A., Xia, L., Duchowski, A., Bailey, R., Holmqvist, K., & Jain, E. (2019). Differential privacy for eye-tracking data. Proceedings of the 11th ACM Symposium on Eye Tracking Research & Applications. doi:10.1145/3314111.3319823</ref> Eye tracking maps have especially unique data because it is often involuntary and undisguisable. One major concern is eye trackers ability to distinguish age from how eyes saccade during a scan path. Further, fixation duration and pupil dilation can reveal gender, race, and sexual preference based off user preferences observed in the scan paths. Aggregation of scan path data can personally identify a person without concrete personal identifiers.<ref name = "identifiers">Liebling, D. J., & Preibusch, S. (2014). Privacy considerations for a Pervasive Eye Tracking World. Proceedings of the 2014 ACM International Joint Conference on Pervasive and Ubiquitous Computing: Adjunct Publication. doi:10.1145/2638728.2641688</ref> | + | The early stages of eye tracking had very little regard for privacy concerns. Since the data was used primarily to create scientific inferences, the data was seen as an open tool to researchers. The benefits to the researchers in the fields of cognitive science, education, and health often outweighed to privacy concerns associated with access to data sets. However, until this point very few methods have been proposed. <ref name = "privacy">Liu, A., Xia, L., Duchowski, A., Bailey, R., Holmqvist, K., & Jain, E. (2019). Differential privacy for eye-tracking data. Proceedings of the 11th ACM Symposium on Eye Tracking Research & Applications. doi:10.1145/3314111.3319823</ref> Eye tracking maps have especially unique data because it is often involuntary and undisguisable. One major concern is eye trackers ability to distinguish age from how eyes saccade during a scan path. Further, fixation duration and pupil dilation can reveal gender, race, and sexual preference based off user preferences observed in the scan paths. Aggregation of scan path data can personally identify a person in a gaze fingerprint, without concrete personal identifiers. The possible breach of this data that links interest and health status threatens to breach the United States' policy of actionable privacy protection. However, since much of the concern regarding privacy concerns of eye tracking is unknown to the public, it is in the hands of developers to make these changes before policymakers. <ref name = "identifiers">Liebling, D. J., & Preibusch, S. (2014). Privacy considerations for a Pervasive Eye Tracking World. Proceedings of the 2014 ACM International Joint Conference on Pervasive and Ubiquitous Computing: Adjunct Publication. doi:10.1145/2638728.2641688</ref> |
| − | + | ||
| + | Many researchers are pushing for adjustments in differential privacy. Differential privacy can be summarized as the inability to be accurately identified even when all other data could possibly be exposed. The perspective of this largely lands in the fear of company to company data sharing, as well as risk of data breaches. The first step to differential privacy is to remove and blatant references to a persons identity. This starts in the form of the name of the participant and any numerical personal identifiers. However, the privacy concerns with eye tracking persists because with the abundance of data sets currently available, and/or accessible to hackers, it has become increasingly easier to link data sets and identify participants. Thus, computer scientists have suggested that privacy concerns in the case of eye tracking go one step further. They suggest eye tracking technology should be matched with machine learning algorithms to add noise and aggregated gaze maps in order to encode the data.<ref name = "privacy"></ref> Other possible solutions include... | ||
{| class="wikitable" | {| class="wikitable" | ||
|- | |- | ||
| − | ! | + | ! solution !! description |
|- | |- | ||
| − | | ''' | + | | '''Affordances of self introspection''' || |
|- | |- | ||
| − | | ''' | + | | '''Levels of abstraction''' || |
|- | |- | ||
| − | | ''' | + | | '''Fuzzing''' || |
| − | |} | + | |- |
| + | | '''Physical Barriers''' || | ||
| + | |- | ||
| + | | '''Policy''' || This suggests that policy makers recognize eye tracking as a form of biometric data. | ||
| + | |- | ||
| + | | '''Tracking indicators''' || This is in the category of the LED light that blinks on to notify the user that their videocamera is on. This privacy control suggests that a light or an icon appears on a screen when eye tracking is in progress. This solves the issue of user information self-determination and consent. | ||
| + | |- | ||
| + | } | ||
Revision as of 18:34, 11 February 2022
Contents
Eye Tracking
Eye tracking is an umbrella term for modern technology that computes eye movements into a data stream. This data provides insights on people's preferences, attention, and interactions. These data streams typically measure the point of gaze and/or the motion of an eye and are matched with behavioral and cognitive sciences to transform this data into usable research. Specifically, eye tracking can collect where a user looks, how long they look, and their overall eye path movement.
The internal processes that are able to capture eye tracking are called fixation and saccades. Fixation refers to pauses in eye movements and are typically very brief. These pauses are registered in our foveal vision and registers with clarity and detail in our brains. This is what eye trackers document as our brains attention is primarily in these fixations. Saccades refers to the rapid movement between fixations. Much of today's research is examining what these patterns and pauses can mean. [1]
location
This is the most basic form of understanding the users' attention. By tracking the location of users' gaze researchers can create a coordinate plot of each fixation to pinpoint each location of interest on the application/ webpage being viewed. However, not all fixations are processed with the fully cognitive mind. Thus clustered fixations on a location are a better interpreter of attention[1]
duration
This tracks the length of time for fixations. Fixations are typically very short and often must be measured in milliseconds. Additionally, fixations are not a reliable source of wanted attention, as they can also represent confusion or signs of zone out.[1]
movement
This tracks the saccades of the eyes. Movements can show the sequence of how users view the screen and/or stimuli. This often helps user experience researchers understand the construction of their pages [1]
Eye tracking technologies have seen recent growth in marketing, advertisement, and other business intelligence improvements. However, these technologies have also been a point of reference for studies on information processing, patterns of decision making, and other topics spanning the cognitive science field.
History
The history of eye tracking dates back to 1879 when Louise Emile Javal, a French ophthalmologist, noticed patients did not read smoothly across pages. This marked one of the first documented scientific interests in eye movement technologies in the form of pauses and fluidity of eye movements during reading.
In 1908, Edmund Burke Huey, a contributor to the field of psychology, built a sense eye tracking device. This device consisted if a contact lens with a small hole in the pupil attached to a pointer. This pointer tracked eye movements by changing position with the movement of the eyes.
The first recordings of eye tracking were published in 1937 by psychologist Guy Thomas Buswell. The recording documented light beams on readers eyes and further documented the relationship between eye movements and reading. [2]
Introduction of eye tracking to modern technology
In the late 1990s, large advertising agencies implemented eye tracking technology to study users' reactions to internet content and to measure the potential of the (at the time) emerging internet market. The current research shift involving eye tracking technology has encompassed a focus on human computer interaction. This field of research has come into demand largely due to demand by marketing strategies for advertisement and big tech companies. This research aims to understand how our brain interacts with the media or web pages we view on our devices. Eye tracking technology has been used in a diagnostic role in this research to understand users' preferences, behaviors, and opinions. This research allows for companies' innovation of strategic user experience designs. Innovation in this field, specifically, has been significant due to its ability to inform the product design process for large companies. [3]
Current Market
The current market for eye tracking has been dominated by the possible introduction of biometric systems in defense organizations. The eye tracking market has very few corporate players, however, there are several systems that currently divide the corporate market: head-mounted vision trackers, optical trackers, and electrooculography systems. Head-mounted systems are typically used in healthcare and research due to its invasive nature, while the other two techniques can be used on a wider corporate scale. [4]
Current Techniques
With modern advancements in modern eye tracking technology, there are currently two main categories, comprised of three to four total techniques, that are utilized in eye movement measurements. The two major categories are measuring the position of the eye relative to the head and measuring the point of the eye relative to its space (typically used for user experience research). [5]
| Method | description |
|---|---|
| Electro-OculoGraphyl | This method measures the electric potential difference of the skin in relation to electrodes placed around the eyes. This method relies on eye position in relation to head position. |
| scleral contact lens/search coil | This is one of the most precise eye tracking methods. This is done through a contact lens with a mechanical or optical reference point. This contact lens typically goes over both the cornea and sclera. This method measures eye position in relation to the head. |
| Photo or Video-OculoGraphy | This method measures the eye under rotation/translation, shape of the pupil, position of the limbus, and corneal reflection. |
Scientific Uses
Eye tracking technology grew in the 2000s uses expanded to implement eye tracking to aid scientific/ medical technologies. Eye tracking technology was created to aid disabled persons communication, identify early ways to identify/diagnose eye abnormalities, and explain cognitive growth throughout childhood. Further, it was used in the fields of computer science to test user experience/ usability of new websites and software.
One of the most prominent uses of eye tracking technology is in the user experience field. The modern use of eye tracking in user experience is by corneal reflection which enables researchers to track eye movements. Corneal reflection requires a light source to shine onto the cornea and a high resolution camera that documents the reflection. The first documented use of corneal reflection was in 1901 and later in the 1950s contact lenses were developed to aid this technique that relied on physical contact with the eyeball. Today, less invasive methods are used and detailed algorithms connect the reflection recordings to point of gaze and make meaning of the eye movements data streams.
Ethical Concerns overview
One major concern of eye tracking technology are consumer privacy breaches. These consumer privacy breaches revolve around user consent in the case of videotaping/camera use in order to obtain the eye tracking data. There have been many efforts in the eye tracking field to de-identify consumer data, as well as create heat or noise maps for tracking rather than direct video footage. [6] However, in addition to big tech companies access to footage without acknowledged consent, eye tracking technology may also reveal private information about the consumers, such as race, gender, etc. Further, these companies typically create decision making models based off the consumers data which reveal intention, cognitive ability, opinion, etc. This has caused concern among the public which has caused a recent pressure on policy makers and corporations to reevaluate consumer privacy regulations. These same concerns are often seen in the implementation of facial recognition technology.
Ethics of Surveillance
Both facial recognition and eye tracking technology have introduced the possibility of a new wave of surveillance technology. There is a concern that the modes in which these technologies allow for one group to track to movements (eye or general) of another group, that this will allow for mass surveillance and a change in authority dynamics due to the power of this tracking technology. Although currently the surveillance used by eye tracking technology has been utilized mainly by retailer and marketing firms use of the data, there is growing concern about the technologies expansion into policing and criminal justice world: threatening a possible surveillance state.[7]
Privacy Ethics
The early stages of eye tracking had very little regard for privacy concerns. Since the data was used primarily to create scientific inferences, the data was seen as an open tool to researchers. The benefits to the researchers in the fields of cognitive science, education, and health often outweighed to privacy concerns associated with access to data sets. However, until this point very few methods have been proposed. [8] Eye tracking maps have especially unique data because it is often involuntary and undisguisable. One major concern is eye trackers ability to distinguish age from how eyes saccade during a scan path. Further, fixation duration and pupil dilation can reveal gender, race, and sexual preference based off user preferences observed in the scan paths. Aggregation of scan path data can personally identify a person in a gaze fingerprint, without concrete personal identifiers. The possible breach of this data that links interest and health status threatens to breach the United States' policy of actionable privacy protection. However, since much of the concern regarding privacy concerns of eye tracking is unknown to the public, it is in the hands of developers to make these changes before policymakers. [9]
Many researchers are pushing for adjustments in differential privacy. Differential privacy can be summarized as the inability to be accurately identified even when all other data could possibly be exposed. The perspective of this largely lands in the fear of company to company data sharing, as well as risk of data breaches. The first step to differential privacy is to remove and blatant references to a persons identity. This starts in the form of the name of the participant and any numerical personal identifiers. However, the privacy concerns with eye tracking persists because with the abundance of data sets currently available, and/or accessible to hackers, it has become increasingly easier to link data sets and identify participants. Thus, computer scientists have suggested that privacy concerns in the case of eye tracking go one step further. They suggest eye tracking technology should be matched with machine learning algorithms to add noise and aggregated gaze maps in order to encode the data.[8] Other possible solutions include...
| solution | description |
|---|---|
| Affordances of self introspection | |
| Levels of abstraction | |
| Fuzzing | |
| Physical Barriers | |
| Policy | This suggests that policy makers recognize eye tracking as a form of biometric data. |
| Tracking indicators | This is in the category of the LED light that blinks on to notify the user that their videocamera is on. This privacy control suggests that a light or an icon appears on a screen when eye tracking is in progress. This solves the issue of user information self-determination and consent. |
- ↑ 1.0 1.1 1.2 1.3 Schall, Andrew ; Romano Bergstrom, Jennifer. (n.d.). 1 - Introduction to Eye Tracking. Eye Tracking in User Experience Design. Elsevier Inc.
- ↑ W., Leggett, D., & Officer, C. (n.d.). A brief history of eye-tracking: Ux booth. Retrieved February 10, 2022, from https://www.uxbooth.com/articles/a-brief-history-of-eye-tracking/
- ↑ Sareen, S. S. (2014). Analysis of eye tracking data obtained by customers' product evaluations (Order No. 1565246). Available from Dissertations & Theses @ CIC Institutions; ProQuest Dissertations & Theses Global. (1616587578). Retrieved from https://proxy.lib.umich.edu/login?url=https://www.proquest.com/dissertations-theses/analysis-eye-tracking-data-obtained-customers/docview/1616587578/se-2?accountid=14667
- ↑ 4.0 4.1 Harwood, T., & Jones, M. (2013). Mobile eye-tracking in retail research. Current Trends in Eye Tracking Research, 183-199. doi:10.1007/978-3-319-02868-2_14
- ↑ Duchowski, Andrew. (2007). Eye Tracking Methodology Theory and Practice. London: Springer London : Imprint: Springer.
- ↑ Liu, Ao ; Xia, Lirong ; Duchowski, Andrew ; Bailey, Reynold ; Holmqvist, Kenneth ; Jain, Eakta. (n.d.). Differential privacy for eye-tracking data. Proceedings of the 11th ACM Symposium on eye tracking research & applications. ACM.
- ↑ Levy, K., & Barocas, S. (2018). Privacy at the Margins| Refractive Surveillance: Monitoring Customers to Manage Workers. International Journal Of Communication, 12, 23. Retrieved from https://ijoc.org/index.php/ijoc/article/view/7041
- ↑ 8.0 8.1 Liu, A., Xia, L., Duchowski, A., Bailey, R., Holmqvist, K., & Jain, E. (2019). Differential privacy for eye-tracking data. Proceedings of the 11th ACM Symposium on Eye Tracking Research & Applications. doi:10.1145/3314111.3319823
- ↑ Liebling, D. J., & Preibusch, S. (2014). Privacy considerations for a Pervasive Eye Tracking World. Proceedings of the 2014 ACM International Joint Conference on Pervasive and Ubiquitous Computing: Adjunct Publication. doi:10.1145/2638728.2641688
- ↑ Rogers, S. (2019, February 05). Seven reasons why eye-tracking will fundamentally change VR. Retrieved February 10, 2022, from https://www.forbes.com/sites/solrogers/2019/02/05/seven-reasons-why-eye-tracking-will-fundamentally-change-vr/?sh=632c7ebd3459
- ↑ Safavi, Seyedmostafa & Shukur, Zarina. (2014). Improving Google glass security and privacy by changing the physical and software structure. Life Sciences. 11.
