We know the secret of your success
Building a C# Gaze Tracking Application using Tobii Eye Tracker to enhance User interaction via Gaze Awareness and Gaze Activatables
Gaze tracking is a concept that is used to track the movements of the eye-gaze to perform specific actions such as select, delete, create, send, and so on. This concept is used in eye trackers. Currently, there is a lot of research conducted to improve the reliability and accuracy of the eye trackers.
Moreover, eye trackers are mostly used for different research areas such as usability research, medical research, and so on. The primary application of gaze tracking technology is used in e-learning and other reasonably agree applications are with BCI, assistive technology, psychology, neuroscience and gaming. However, this project identifies that eye-gaze tracking technology can be used for day-to-day activities such as email, chat and e-commerce.
This project identified that the functions of Tobii eye tracker could be divided into two significant actions which are gaze awareness and gaze activatable. The gaze awareness action is mainly by performing activities from gaze positioning and the gaze movement. On the other hand, gaze activatable is performed using the gaze position combined with the keyboard press or mouse click. Moreover, this project develops a system using C# and Tobii eye tracker to perform functions such as using gaze awareness (gaze position) to change the colour of the button, opening of form by clicking a button using gaze awareness (gaze position and the delay), using gaze awareness (gaze position) to verify whether the user is presented or not and whether the gaze tracking is on or not, using the gaze activatable (gaze position and press keyboard shift) to open more quote in the form user interface, using the gaze activatable (gaze position and press keyboard shift) to close the quote in the form user interface and using the gaze activatable (gaze position and press keyboard shift) to press the close the button.
Keywords: gaze tracking, Tobii, eye tracker, Human-Computer Interaction (HCI)
In general eye gaze is described as a line of sight of an individual (Zhu and Ji, 2007). It is considered that eye gaze is an individual’s focus of attention (Zhu and Ji, 2005). The gaze tracking is active research for several decades due to its possible applications in different areas which include Eye Disease Diagnosis (EDD), Human-Computer Interaction (HCI), Human Behaviour Study (HBS) among others. In the early research regarding gaze tracking, the gaze trackers were mostly disturbed because the tracker requires physical contact with the user which includes placing a white dot within the eyeballs (Milekic, 2003) or attach electroplate near the eye (Hyoki et al., 1998). Additionally, it was also required that the user of the gaze tracker should not move their head which triggered discomfort during the usage. However, due to the technology advancement current gaze tackers uses microcomputers and video cameras which ease the usage (Zhu and Ji, 2005). Therefore, this project focuses on the gaze tracking technology to support multi-touch interaction and for everyday tasks.
The project aims to explore the potential of gaze tracking to support multi-touch interaction and for everyday tasks which includes email, chats and e-commerce.
The objectives of the project are:
Objective 1: Design a catalogue of interactions supported by gaze tracking.
Objective 2: Different test prototypes by the in-lab experiments with users to identify which interaction prototype design increases the usability of the applications.
Objective 3: Make use of the iterative, user-centred methodology to capture the buyer and the client perspective
1.3. Research Questions
The research question of this project is
“Can gaze tracking become an effective supporting technology in everyday tasks of an individual?”
Gaze tracking technologies were introduced several years ago. However, the technology did not become common because of its high cost. For example, the price of high-end gaze trackers is starting from 1000 pounds. In general, the technology is used to analyse the usage patterns of different applications which include e-commerce sites, general user interaction websites and forms. Though, the gaze tracking technology was not used for user interaction with the applications because the designers of the application could not expect the general users to won gaze trackers.
Nowadays, the cost of the gaze tracking technologies is reduced, and they became consumer-grade products. The reduce in the cost widen the opportunity to use gaze tracking as a support input modality to interact with computers. However, gaze tracking will only become a mainstream technology for interaction, if they prove to be useful for the users of the application.
Therefore, this project going to identify how the gaze tracking technology is used for different applications.
This project identifies how gaze technology is useful for multi-touch interaction and for everyday tasks which include email, chats and e-commerce. Moreover, it also implements some simple prototypes to validate the findings. However, this project will not implement complex prototypes.
The detail description of the structure of the report is provided below:
Chapter 2 provides a detailed finding of the background of the participants regarding their knowledge about eye tracking, and the secondary research regarding the eye-tracking.
Chapter 3 describes the methodology used for this project which includes quantitative research to identify whether people are aware of the eye tracker and then the approach used to create the catalogue of the usage and drawback of Tobii EyeX tracker and the approach took in implementing the technology demonstration and testing them.
Chapter 4 provides the complete information of the catalogue created for the Tobii EyeX tracker
Chapter 5 explains the implementation of the system using the Tobii EyeX tracker
Chapter 6 provides a clear output of the testing of the system implemented
Chapter 7 provides the user evaluation performed on the system implemented
Chapter 8 provides the recommendation on how the Tobii EyeX can be used in platforms such as PayPal.
Chapter 8 provides a conclusion and future work of this project
This chapter mainly focus on providing the detailed information regarding the findings of the background knowledge about the participants regarding the eye tracker before the experiment and the secondary research regarding the gaze tracking and eye tracker (Potential market awareness). The methodology used for the potential market awareness study and the secondary research are provided in section 5.3 and 2.3.
See chapter 4 for the complete analysis of the interview and survey conducted for the potential market awareness. However, from the findings it is reasonable to say that majority of the people are aware of the use of eye trackers and agree that they will be able to use the eye tracker in their day-to-day activities if it is affordable.
A human being’s area of fascination or present line of sight is controlled by gaze. The connection between the surface of the observed object (e.g. monitor) and line of view is characterized by the fascination point (Popien et al.,2015). With the use of gaze, a user’s plan to communicate without command, or to allow (fascination reliance) accommodation and vibrant intensity of focus may be deciphered. The probable advantages of including the movements of the eye into human and computers communication are many (Hansen, 2016). Let’s say, for instance, a computer can decipher a user’s wish and most likely enable a computer to establish certain intellectual conditions of the user like exhaustion or confusion by merely knowing the location of a user’s gaze (Vicente, 2015).
The direction of eye gaze has the capacity of demonstrating a user’s interests; it is a possible window into the present intellectual practices. Interactions through the eyes’ direction are the fastest means of communication amongst humans. Furthermore, instantaneous observation of gaze location allows the launch of changes in display that depend on the spatial or temporal features of the movements of the eye (Swaminathan, and Ramachandran, 2014). This type of technique is called the gaze-contingent display paradigm. For instance, a person’s fixation on the monitor which can consequently insinuate a user’s area of interest can be determined by his/her gaze. Suitable actions such as increasing resolution or size of the area that preoccupies a user’s gaze can thereafter be taken. Another instance is reducing the size on bandwidth by placing high-resolution data exclusively in the location a user is keen on (Lai et al., 2015).
Gadgets that have the capacity of to gauge a person’s direction of gaze is called Eye gaze trackers (EGTs) (Young and Sheena, 1975). Some latest study can be seen in Glenstrup and Engell-Nielsen (1995) and Duchowski (2007)
In the early times EGTs were created for the scientific discovery in monitored conditions or laboratories. Information derived from eye gaze has been utilized in neurology, ophthalmology, psychology, and similar areas to review features of oculomotor and faults, and their connection to intellect and psychological conditions.in recent times, more applications to review marketing and advertising, and also human considerations engineering to assess computer web sites and interfaces are available even though they are as at yet restricted to laboratories only. EGTs for a long time have been alluded as input gadgets for computer interfaces, profitable efforts are as at yet, restricted to military applications and the creation of interfaces for disabled people (Boraston, and Blakemore, 2007).
A picture of the eye is dealt with in three major steps in standard gaze trackers Firstly, the specular mirror image of a fixed source of light is seen in the picture seen by the eye. Secondly, the centre of the pupil is discovered. Lastly, the comparative location of the mirror image of the light to the centre of the pupil is calculated. The location of the gaze direction is ascertained from the data about the comparative locations as illustrated in Figure 1. The user needs to be still or put on a unique headgear in order to retain a regular offset between the eye and the camera’s position in most of the present gaze tracker systems.
Thus, the use of Gaze tracking is imperative for smart graphics and human–computer interaction (HCI) (Poole, and Ball, 2006). Many methods of which commercial eye gaze trackers are inclusive have been created. These can be mainly divided into video-based methods and non-video-based methods. Normally, non-video-based techniques utilize certain unique contacting gadgets connected to a user’s skin or eye to get his/her gaze. In this way, they are meddlesome and meddle with the user. An example can be seen when a user’s skin, around the socket of his/her eyes have electrodes attached to it to measure variations in the orientation of the likely difference between cornea and retina (Fogt, 2013). The day to day use of this method is too problematic as a close connection of electrodes to the user is required. In addition, a user’s eye has non-slipping contact lens placed in front of it. In as much as this method helps to get the direction of gaze accurately, it however is very awkward and discomforting and therefore makes nonlaboratory tasks impossible.
Duchowski (2007) postulated that classifications of eye gaze tracking applications are interactive or diagnostic. Interactive applications utilize information gotten from eye gaze to react to or communicate the user as a result of movements of the eye. Diagnostic applications utilize information gotten from eye gaze as measurable proof of the users graphic and attentional procedures. A lot of conventional methods for eye gaze tracking are meddlesome. That is, the user has to have physical equipment attached to him/her. Electrodes, contact lenses and head mounted gadgets are examples of wearables included in these methods. Non-meddling methods (or remote methods) are predominantly based on graphics, i.e., cameras are utilized to capture images of the eye. Certain camera-based methods may be to some degree meddlesome if it is necessary for them to be worn on the head. For a couple of interactive applications, a user heavily relies on eye tracker to carry out certain tasks (i.e. limited options or no alternative gadget) Likewise for diagnostic applications, eye information can be captured in the process of a short experiment and processed later, the time needed to organize the eye gaze tracker and the uneasiness that is the likely outcome of the equipment do not pose a problem (generally speaking).
Remote eye gaze tracker (REGT) is more comfortable to use. It is more user friendly and setting it up is faster (Frazier et al., 2016). The use of the method is more durable compared with intrusive method. Its precision is however lower than intrusive EGTs but on long term use, they are more suitable. The pupil–corneal reflection method is usually publicized as a remote gaze tracking technique which is vigorous to certain head movements and easily adjusted. Majority of these claims are regrettably false. Their precision is substantially influenced by the small head movement endured by such gadgets, and a such, trials are normally done utilizing bite bar or chin rest to limit head movements, which is very discomforting to the user. According to Schnipke and Todd (2000), it is difficult to gather trustworthy eye tracking information utilizing a commercial REGT technique. In setting up a popular software application, to experiment its usability, an experienced operator in eye tracking could not make the technique track 10 out of 16 subjects, and none of them wore glasses (Singh et al., 2016). They also face the challenge associated with difficulty in setting up the method, such as adjusting the brightness settings and placing the camera in a way that there are no interferences from eyelashes, and also the underlying challenges of the method associated with adjustments, glasses, dry eyes and pauses of the method in cases of head movements (Kowsari et al., 2014). Eye gaze-based interactive application has prospects to modernise the way we use computers regardless of its shortcomings (Kassner et al, 2014). Lewis examines design subjects for building smart agent-based user interfaces (Manojlovich et al., 2003). Agent is used to describe the computerized features of human–computer interaction (HCI). These types of interfaces have the capacity to predict commands or carry out tasks independently. Before they can do that, they need dependable ways to identify the user’s focus of attention to imply the users ‘‘intention.’’ Vertegaal expresses design challenges for a similar interaction paradigm called attentive interfaces. It illustrates the benefits it could derive from REGTs. Zhai (2003), Glenstrup and Engell-Nielson (1995) and Duchowski (2007) have postulated other applications of eye trackers.
Gaze tracking applications are classified as either interactive or diagnostic. Interactive applications utilize eye movements instead of an input means like mouse, thus, user communicates with a computer with the eyes only. The diagnostic type captures the learner’s focal point, consequently, showing proof of what attracts a learner within a period. (Al-Khalifa and George, 2010).
Tracking learner's feeling
An individual’s centre of attraction and interest level are shown in the information gathered from gaze-tracking devices. Getting information about stress, focal point, problem solving, leisure, fatigue, learning successfulness and more is possible as a result of eye position tracking and indirect means like blink rate, fixation numbers, gaze position and duration. (Al-Khalifa and George, 2010).
If for instance, the average pupil size has increased continuously within some period, a user’s workload may have also increased. A reduced blink rate in the same time would additionally affirm such assumption (Magnussen et al., 2017). Such proof when discovered, could for instance, be utilized in powerfully altering the learning path, suggesting a subject related to the major subject but less complicated. Additional data may be displayed if per chance, the user is possibly having difficulty in comprehending something (Ivanovi? et al., 2017). Not undermining the impacts of several external factors, nonetheless, it is difficult to be certain beyond a shadow of doubt that these signs are gotten from a user’s passionate state. Subsequently, the system can indirectly help the user with gradual assistance instead of undertaking direct actions like displaying help windows (Magnussen et al., 2017).
For example, upon discovery of signals of high mental workload or of non-comprehension, the system simply suggests links to supplementary material, which increasingly broadens as stress signs persist. Once eye data recommend that the user may be exhausted, and the session has lasted for more than a configurable time interval, one hour for instance, a message encouraging to take a break appears (Magnussen et al., 2017). Below are some experimental proofs in physiology/psychology:
- The instability of the rhythm of the pupil area affects mental workload.
- Increased complexity of the task reduces Saccade length and occurrence rate.
- Increasing fatigue reduces Saccadic and blink velocity.
Tracking learner's interest
Learner's interest can be tracked following his eye movement on a web-based learning platform. In this example, "Gazetracker 2.0 Beta" was used to find a learner’s areas of interest on a web page that comprises of a chapter of an algorithmic course (El Haddioui and Khaldi, 2012). The "Eye mouse" option was enabled to redirect the mouse cursor to the gaze position (Pfeuffer et al., 2015). Any eye tracker that supports the “Eye mouse” can be used (Pfeuffer et al., 2015).
Tracking information will be stored in the database; gaze duration statistics of every area and learner can be found by teachers. Durations are converted from millisecond to second for easier comprehension.
This research attempted to utilize eye movements to imply the significance of records in the recovery phase of a data search. The scholars associated pertinent decisions to increments in pupil width, as a wider width usually indicates high enthusiasm in the subject matter (Pan et al., 2004).
This exploration tries to get a more thorough comprehension of the investigator’s actions and study before choosing a web-based document. Visual files empower us to figure out the abstracts been viewed and read by a user and also duration and order of the abstracts. All through the historical backdrop of gaze tracking study, a few essential factors have risen as critical pointers of visual practices, involving scan paths, fixations, pupil dilations and saccades. Spatially steady gaze lasting for roughly 200-300 milliseconds, in which ocular attention is pointed to a definite area of the graphic show is known as eye fixations (Kim et al., 2015). Fixations symbolize the occasions in which data procurement and processing can happen, and consequently, fixations were the files most applicable to this present assessment. Pupil dilation is usually utilized as a measure to check a person’s interest or excitement in what they are seeing
User’s experience can be improved on websites through numerous methods. Social presence inclusion is one of them. The theory of Social presence defines the consequence or impression of the presence of another person when a user is communicating with a definite means. This hypothesis was originally applied to and explored utilizing telephony only; nevertheless, study shows other communication technologies can also apply this theory
Social presence inclusion might be very significant on e-commerce websites since they are more automated and impersonal than conventional one-on-one commerce, they may be seen as lacking human affection. Social gestures like pictures, can build positive impressions of group interaction (e.g. online discussions) and warm sentiments, resulting a more enjoyable experience of a user. A current research demonstrates that pages that incorporate human faces are seen as having a higher level of social presence (Cyr et al 2009). A similar report demonstrates that users discover pictures that incorporate human faces as more fascinating than those pictures that do not incorporate faces (for instance, those that show human body) and those pictures that do not have individuals in it.
Eye tracking likewise supported the studies that demonstrate users gaze is often more intense on faces on a picture (djamasbi et al 2008). These discoveries allude that facial image inclusion on a web page will probably influence how visually alluring users see the page.
study also recommends that users are more prone to trust a page that is visually alluring. This demonstrates that visual interest is probably going to be a good indicator of trust. To analyse this clearly, we conduct investigations to decide whether for sure there is a fundamental connection between visual interest and trust. Current eye-tracking research recommend that accessing a user’s gaze can help to have a better comprehension of a user’s responses. Eye tracking has been noted as especially helpful when evaluating the options of a design or the usefulness of a specific part on the page. Consequently, this study focuses on the evaluation of users’ fixation so as to explicate their behaviour (Djamasbi, 2010).
The outcome of this study demonstrate that users found a page with pictures of people more visually enticing than a page with pictures of non-people (“logos”) and that their visual appeal ranking was essentially related to their trust rankings. Moreover, our evaluation demonstrates that partakers finished tasks utilizing the Faces model crucially faster than with the Logos model. More so, the horizontal order of data on the web page significantly affected the completion time (Wang et al., 2014).
The discovery that visual interest is a noteworthy indicator of trust is of specific interest to organizations. These outcomes demonstrate that, by enhancing the interest of their web pages, organizations can assist to enhance confidence in the web page and, probably, in the organization itself. Our outcomes show that, when it is suitable to the content, the picture inclusion of faces can help to increase the attraction to a page (Djamasbi et al., 2010).
Recommendations from the heat maps show that pictures of faces drew more attention to textual data which filled in as an anchor for partakers, as they could have a higher fixation length on the important information as opposed to the numerous fixations and reduce fixation length found on the Logos model. Together, these outcomes recommend that organizations can profit from including faces to their websites when their inclusion suitably describes message communicated.
Moreover, the outcomes demonstrating a quicker completion time for the task applicable to the left side of the page recommend that involving repeatedly finished tasks on the left side of a page can possibly increase the efficacy of users and, therefore, enhance the user encounter.
This study has fundamental hypothetical and practical consequences. Hypothetically, it increases past study on the impact of pictures, and specifically human pictures (e.g., faces), on user conduct. It likewise gives proof of a positive connection between visual interest and trust. Additionally, this research adds to human computer interaction study by giving additional proof that fixation patterns can give significant data in regard to user behaviour. This actual measure gives exceptional value as it constantly captures user conduct, though personal measures normally happen towards the start and/or end of sessions. This utilization of eye tracking can inspire a new and productive line of study.
Essentially, this examination gives discernment into the user encounter, empowering organizations to enhance the web experience of their users. Such a development can help organization to increase their reach and, hence, their market share. Specifically, the outcomes related to areas of web pages situated underneath the fold. Since users display a top down viewing inclination, tempting users to watch content situated at the base of web pages, especially beneath the fold, is often reported to be difficult. Therefore, discovering approaches to draw attention to components beneath the fold is of particular value to organizations
A wide range of strategies have been utilized in tracking the direction of the eye since the inception of eye tracking technology in reading study more than 100 years ago (Poole and Ball, 2005):
Electro-oculographic approach (EOG): Fenn & Hursh formed the EOG (Itsuki et al., 2993). It is an old technique used to calculate the contrasts in electric potential caused by eye movement. Electrodes captured these electrical potentials placed around the eyes. It is hard to know the actual eye position even though the calculation has a precise time resolution, and as such, it is not very valuable in HCI (Human-Computer Interaction) investigation (Baccino and Colombi, 2001).
Scleral search coils approach: Putting on large contact lenses which covered the cornea (a transparent membrane securing the front of the eye) and also sclera (the white part of the eye) is needed in this method (Poole & Ball, 2005). The concept is to generate a magnetic field and to recognize within this field the different electrical sign using a special lens put on the user's eye. The user is kept in a magnetic field made by three horizontally organized coil, either vertical or lateral position. These three features marked the stare position. This method is a very exact technique but at the same time, stressful (an ophthalmologist has to be available at every execution) and to a large extent, this seriously confines its utilization for mental or biotechnological reasons (Baccino & Colombi, 2000).
Corneal-reflection approach: Most commercial eye-tracking methods accessible today calculate the “corneal-reflection/pupil-centre” approach with point-of-regard. These types of trackers mostly comprise of a standard computer with an infrared camera placed under a monitor, with a software that processes image to discover and recognize the eye characteristics used for tracking. In execution, there is first of all pointing of an installation of infrared light from an LED (Light-Emitting Diode) in the infrared camera into the eye to make strong reflections in target eye characteristics for easier tracking (infrared light is utilised to prevent astonishing the user with noticeable light) (Poole & Ball, 2005).
Figure 1: Corneal reflection and bright pupil (Poole & Ball, 2005).
The figure 1 thus demonstrates the infrared light reflections that enters the retina and a huge amount of it is reflected back, making the pupil look like a vivid, well-defined disc (called the bright pupil effect). Infrared light in likewise manner creates a little, but sharp, glint known as corneal reflection (or first Purkinje image). As soon as the software that processes image has recognized the centre of the pupil and the position of the corneal reflection, the vector in between them is calculated, and, with more trigonometric calculations, the discovery of point-of-regard is realized. Despite the fact that the realization of approximate point-of-regard by corneal reflection only by tracking both features eye movements is possible, basically, be dissociated from movements of the head (Poole & Ball, 2005).
These two features of the camera image are discovered by the eye tracking system and decide the location of the user’s gaze based upon the both features’ separation.
Eye-Movement Metrics
Movement of the eye is normally classified as fixations and saccades; fixation describes the somewhat static position of the eye, collecting or encoding data, and saccade is the movement of the eye happening amid fixations, usually ongoing for 20 to 35 milliseconds. Most saccades’ objective is taking eye movement to the following viewing spot. Graphics processing stifles instantly throughout saccades to prevent graphic image distortion. Majority of the eye data is accessible throughout a fixation and not throughout a saccade (Shrestha and Owens, 2009).
Scanpath is the outcome of successions of fixations and saccades. They are valuable in examining intellectual salience, attraction and intent. Other biological elements such as gender can also influence scanpath. Gaze tracking in human-computer relationship investigation usually examines the scan path with the aim of usability, or as an input technique in gaze-contingent illustrations, likewise called gaze-based interfaces.
The table below illustrates some eye-tracking pointers that are valuable in assessing web applications usability. There are two kinds of pointers; pointers of exploration of data that demonstrate the spatial traces of the movement of the eye and pointers that show the data signs to imply the moments of processing (Baccino & Colombi, 2000).
Gaze tracking vs eye tracking
Gaze trackers essentially calculate the rotation of the eye as regards the calculating technique. Whenever the calculating technique is worn on the head, as with EOG, subsequently, eye-in-head positions are calculated. Whenever the calculating technique is placed on a table, as with camera placed on table (remote) techniques or sclera search coils, then gaze positions are calculated saccade (Shrestha and Owens, 2009).
The head position in several applications is fixed utilising forehead support, a bite bar, or something related so that gaze and eye position are identical. Additional instances allow the head to change position, and movement of the head is calculated with techniques like video-based or magnetic head trackers. In trackers placed on the head, head direction and position are included in the eye-in-head direction to decide the direction of gaze. Table-mounted techniques like search coils deduct gaze direction from head direction to decide eye-in-head position.
The discreet nature of an artificial neural network-based gaze tracker is one of its major advantages; the user can freely move his head. The eye has to be directed to every image frame to account for the moves in the comparative direction of the eye and the camera, the location of the eye has to be in every image frame. The right eye is detected by looking for the mirror-like image of a static light in the user’s face image in the present technique. For the most part, this can be recognised by little bright area enveloped by a black area. The location of the reflection is utilised in restricting the lookout for the eye in the subsequent frame. The eye’s image is detected inside a window after the window enveloping the reflection is removed.
In deciding the organisation of the direction of the user’s gaze, the pixels removed from the window are utilised as artificial neural network inputs. The ANN replicates the forward pass and the reading output unit decides the gaze arrangements. 50 output units organise the output units to stipulate the X arrangement, and the Y coordinate utilises 50 units. X and Y-axis output units utilise a gaussian output symbol, like the one ALVINN independent road trailing technique used (Pomerleau and Baluja, 1993). Gaussian programming symbolises the reaction of the network by a Gaussian-shaped launch height in a trajectory of output units. The height location inside the trajectory symbolises the location of the gaze beside either the X or Y axis.
Information for the education is gathered by teaching the user to graphically trace a cursor on the motion. The movement of the cursor is a predefined channel. The eye image is computerised and put together with the (X, Y) cursor’s coordinates. 2000 image/location pairs are gathered in all. Every network this investigation explained is taught using very similar guidelines for 260 epochs, utilizing typical error back broadcast (Pomerleau and Baluja, 1993).
In conclusion, it is safe to say that gaze tracking technology is a complex research area that is still in progress.
YOU CAN GET THE FULL REPORT WITH CODE HERE
Last updated: Apr 08, 2020 08:13 AM
Your one-stop website for academic resources, tutoring, writing, editing, study abroad application, cv writing & proofreading needs.
