Students’ Research Works – Spring 2017: User Experience and Implicit Feedback subgroup (PeWe.UX)

Usability of Information Visualizations

Adam Bacho
master study, supervised by Róbert Móro

Abstract. Today, people collect data from a vast number of different sources. Their goal is to analyze this data for specific purposes. One of the most common ways of visualizing data in the process of knowledge discovery are decision trees, which may in certain exploratory analysis tools support decision of users. Specifically, we want be able to detect problematic sections of individual users when working with decision trees. The main resource that will enter the process of problems detection will be data obtained by tracking users’ gaze. These provide us with a large number of metrics that can significantly affect the quality of problematic parts detection. Another metrics that we plan to deal with are the number and length of fixations, saccades and AOIS’s semantics as it will be important to find out where the users look mostly, or where problems arise. It will be interesting to include metrics of the outside (ambient) and central (focal) perception as well. All these data are then used for problems detection using random forests, since it is highly effective when used for different purposes.

to the top | to the main

Prediction of users’ personality traits based on task solving on the Web

Veronika Balážová
master study, supervised by Róbert Móro

Abstract. These days people are using web pages and web applications more often than they had used to. How user interacts with web page does not only depends on interface or on usability of page, but it depends on user’s characteristics as well. For example, cognitive abilities or personality of users can influence the way, how they look at web page and how they scan it. Knowing the user’s characteristics can be useful for example in personalization of graphical interface or in marketing ads, which can appear on the page depending on the user personality.

In our work, we plan to analyse impact of the type of the user personality on his way of dealing with typical tasks in specific domain, for example in online shops. Our goal is to predict user personality based on his interaction with the information system and based on eye tracking data.

to the top | to the main

Multiplayer game based on eye tracking

Michal Baňas
bachelor study, supervised by Jozef Tvarožek

Abstract. Computer games are an integral part of today’s society. With the development of modern technologies and games themselves, much attention is devoted to inventing new gaming practices, strategies, or controls. Eye-tracking techniques, however, often stay untapped, and there aren’t many games that can fully use the benefits of these sensors.
The aim of my project is to explore ways to work with dynamic content using a software prototype of a multiplayer web application, that will capture users’ actions in detail and explore the way people work with this content.

to the top | to the main

Analysis of eye movement patterns depending on task in the web environment

Matej Červenka
master study, supervised by Jozef Tvarožek

Abstract. Users perform various types of tasks on the web every day. Tasks includes searching via search engine, shopping, reading, chatting etc.. It is able to increase efficiency of these processes by recording and evaluating interactions of mouse or keyboard. Except of these devices it also able to use eye tracking technology which allow us to better understanding how users are thinking while performing their tasks. Users are thinking different while performing various types of tasks, which reflect in different gaze plots patterns.
In our research we will try to find and automatically identify patterns in gaze plots which are specific to some type of task in web environment. On the basis of these patters we want to predict what kind of task user was doing.

to the top | to the main

Identifying User Characteristics by Eye Tracking Analysis

Mária Dragúňová
master study, supervised by Jozef Tvarožek

Abstract. Human individuailty is composed of many aspects such as knowledge or cognitive abilities. These aspects have impact on every day life and activies. Using computer applications is not an exception. Eye tracking analysis is a good way of identifying and evaluating some of users’ characteristics and knowing the user is a neccessity if we would like to take into consideration his individuality when we create applications or evaluate usability of a user interface.
We propose to identify individual characteritics of user using appropriate eye tracking measures. We would like to create such a program, which can automatically detect distinctive user characteristics.

to the top | to the main

Towards Quantitative Eye-Tracking User Studies of Mobile Applications

Peter Dubec
master study, supervised by Mária Bieliková

Abstract. Usability is a quality attribute that assesses how easy user interfaces are to use. Nowadays usability testing of mobile applications with use of eye-tracking is realized on individual basis and evaluated manually. Therefore in most cases these tests are qualitative and not quantitative because we are able to collect only small amount of data.

Our main goal is to make bulk testing of mobile applications possible. In our work we have created user study method, which thanks to emulation on computer enables more effective qualitative usability testing of mobile applications with use of eye-tracking. It is clear that there are some specific usability problems that can be identified only during testing on real mobile device, but we have discovered and confirmed that we are able to detect multiple usability problems also during emulation on PC. We defined two lists of potential usability problems with mobile applications, where one list contains problems that can be identified with use of emulation and second contains problems that can not be identified with use of emulation. We also focus our work on automation support of user studies evaluation thanks to automatic evaluation of specific metrics.

to the top | to the main

Automatic Web Literacy Detection

Patrik Hlaváč
doctoral study, supervised by Mária Bieliková

Abstract. Usability studies in the web domain are based on various metrics, but the question is how to apply these metrics to evaluate a larger group of people. When we consider that every user has different qualities, skills and experiences, we could expect that the results of testing of same scenarios will be different. We focus our research to show that quantitative studies can provide more accurate results if we work with information about personal characteristics of participants.

Now we deal with web literacy detection from implicit user feedback. We are using three sources for collecting data: browser (mouse, keyboard) and eyetracking data for modelling three lines of interaction. These will be used for pattern finding.

to the top | to the main

Cognitive Load Methods Based on Pupillary Response

Tomáš Juhaniak
master study, supervised by Mária Bieliková

Abstract. Since first successful results in our bachelor researche, we are able to measure cognitive load with only eyetracker-generated data. Our non-intrusive method based on pupillary response and screen recording has potential for simplest but precise approach to cognitive load analyze. In this work we test our method in multiple experiments or scenarios and we examine new methods, just as method working without screencast or “absolutely” precise method based on self-developed polarization glasses. Since these methods can be managed without deep knowledes of eyetracking field, we expect including our method to several UX experiments.

to the top | to the main

User Experience and Emotions

Ondrej Kaščák
bachelor study, supervised by Róbert Móro

Abstract. User experience (UX) along with usability are crucial parts of every software product. In order to achieve the best possible usability and user experience, it is necessary to perform testing with real users who can provide us feedback.

During our experiment we want to use Emotiv EPOC electroencephalograph (EEG), which measures the cerebral activity of a person. The goal is to measure physiological changes of a person evoked by emotions that user experiences during testing.

Main aim of our work is to propose a method of emotion recognition and to verify possibility of using mentioned device and collected physiological data during usability testing. Data collected during test sessions will be further processed, analyzed and we will use machine learning algorithm (SVM) to classify emotions whether they are positive, negative or rather neutral.

Currently we are conducting experiments in UX Lab with participants. Experiment consist of simple web based tasks on website ZlavaDna.sk

to the top | to the main

Detection of user features based on eye-tracking method

Ondrej Kipila
master study, supervised by Jakub Šimko

Abstract. The characteristics of the user are very useful information in information systems. If the information system is able to automatically detect the characteristics of the user (eg. the mood, interest, fatigue, personality characteristics, capacity short-term memory), the system can use this information to adapt its functionality in such a way that as far as possible meets the needs of the user. An example would be automatic content recommendation. Another option is the use of prediction of the user’s activities and achievements in the system. One way of identifying a characteristic of the user uses tracking of user’s operations in the system (e.g. clicks, scrolling, text entry). Recording these activities can be further combined with user’s eye tracking data.

In our work, we plan to predict user intent from his eye-tracking data. Our main goal will be categorizing users by analyzing their eye movement patterns, and predicting their actions on the web page.

to the top | to the main

User modeling using eye-tracking and size of the pupil

Sandra Kostova
master study, supervised by Jakub Šimko

Abstract. Characteristic of the user of the information system can be very valuable information that can be used in improving system’s usability and personalization. Depending on the personality of the user, his mood or interests, his behavior can be predicted. On that basis, we can make design changes to the system or introduce some elements that will enable automatic adjustment and system personalization.

We propose to detect some of the user’s characteristics by using appropriate eye tracking measures. Our main goal will be to predict the curiosity level of the user by analyzing his eye pattern movements. This characteristic is closely connected with other characteristics such as: extrovertism, intelligence, antagonism so we will be also able to predict them as well. The results will be improved by using appropriate personality tests.

to the top | to the main

Automatic evaluating usability of applications with eye-tracking technique

Lenka Kutlíková
master study, supervised by Jakub Šimko

Abstract. In almost each study, participants should read and follow written instructions. During experiments and also in real life users often do not read them. They just skip them and do the task on their own. In case of doing experiments it can cause bias in results. In order to improve the quality of experiment results we have to find participants who do not read and follow instructions. We can ask them. But in the case of not reading instructions they can lie. Better approach is to use eye-tracking technique to catch their gaze. We supposed that eye-tracking metrics indicate if users read and follow instructions.

The main idea of our research is to find people during experiments who do not follow instructions. Eye-tracking technique is used. We designed and conducted an experiment with more than 50 participants. Each participant should do 4 different simple tasks. Tasks are well-known games so participants should be able to do them also without reading instructions. Instructions are designed as very simple game rules with one contradictory instruction. According to this teaser we evaluated manually if participant followed instructions or not. We identified eye-tracking metrics correlating with following instructions. We would like use machine learned model based on these metrics for prediction if people follow given instructions.

to the top | to the main

Predicting Correctness in Human Computation Tasks with the Use of Eye Tracking

Aleš Mäsiar
master study, supervised by Jakub Šimko

Abstract. Human computation is often used for acquiring information, that computers cannot obtain automatically. Many human computation approaches, however, struggle with the problem of correctness validation when processing human input. Several existing methods deal with this problem, but lot of them are limited by insufficiency of gathered data.

This work proposes a method, that deals with the problem of correctness evaluation of individuals answers in a human computation task using also information acquired from gaze tracking device. This work also addresses application of the designed method on specific tasks and its integration to the task providing system for more efficient utilization of human work.

to the top | to the main

Gaze Visualization

Michal Melúch
bachelor study, supervised by Róbert Móro

Abstract. Recent growth of interest in studying eye movements has triggered the development of new methods for common scanpath retrieval. These can be then used for modelling the behavior of a much larger user base. Each of the analysis methods was developed in order to solve a specific problem and their measures differ with respect to their target characteristic. Our work is focused on development of a web application suitable for comparing methods of a common (trending) scanpath retrieval. Primary benefit of such application consists of providing a unique way of evaluating and visualizing scanpath data in a web browser.

to the top | to the main

Visualization of user activity in an interactive application

Lukáš Meňhert
bachelor study, supervised by Mária Bieliková

Abstract. It has been proven, that opening up the the learner model to students may lead to further enhancement of their learning experience. Our project then aims to make the learner model explicit to the users of adaptive learning framework using graphical visualization. For different types of users, there are different types of visualizations that may prove interesting and so we will be visualizing data from two different viewpoints. The first one is the viewpoint of a student whose primary goal may be to get the highest possible score on a test or for us to make his learning process much more enjoyable. Second viewpoint is that of a teacher whose primary goal may be to review the most problematic questions for students and then adjust material explained on the next lecture.
CONTENT START
Analysis of interactions in the domain is then essential to make visualizations, that add value to the learning process both from students and also from teachers perspective.

to the top | to the main

Identification of the User Expertise Based on Patterns in Eyetracking Data

Martin Mokrý
master study, supervised by Róbert Móro

Abstract. Identification of repeating patterns from eye tracking data (patterns like sequences of fixations, saccades or areas of interest) is considered to be an important step in the eye tracking analysis. Its main focuses are: explanation of recorded interaction, comparison of ways, in which different users interact or clustering of users based on their similarity. This can be used in evaluation of recorded interaction (UX testing) or customization of graphical interface based on the identified situation (pattern). This situation, e.g., a systematic scanning of a web page can imply unfamiliarity of a user with a specific web page. On the other hand, reoccurrences of a similar situation can imply activities of a skilled user.

In our work, we focus on automatic identification of patterns in scanpaths in the eye tracking data, related to the domain skill level of the user. The first phase consists of identifying proper features of fixations and saccades (features not task-related and with high frequency of occurrences) and analysis of methods for creating commons scanpaths and quantification of similarities between scanpaths. The goal of the second phase is to implement machine learning model capable of automatic identification of the domain skill level of user using just the first few seconds of the interaction.

to the top | to the main

Inferring User Characteristics from Sensor Data

Róbert Móro
doctoral study, supervised by Mária Bieliková

Abstract. When users interact with applications or on the Web, we can observe individual differences in their behavior. These are (to a large extent) manifestations of their characteristics, such as their interests, knowledge, psychological traits or cognitive abilities. Especially the latter two are traditionally hard to infer with the questionnaires being the most reliable and (often the only) means of acquiring them. However, the increasingly more accessible and affordable sensors make it possible to utilize their outputs as signals for modelling and inferring these characteristics.

In our work, we focus especially on eye tracking and research the novel methods of eye movements analysis. From these, the methods of scanpath analysis are the most prominent; they represent the eye tracking data as a sequence of fixations and aim to discover interesting and recurring patterns within a single scanpath or in a cluster of scanpaths.

Additionally, we examine also the use of other physiological sensors for inferring the user characteristics, such as EEG or GSR that can complement the eye tracking data and provide a more detailed picture of the users’ behavior.

to the top | to the main

Automated evaluation results of user studies

Dániel Papp
master study, supervised by Jakub Šimko

Abstract. Nowadays websites are part of everyday life and people want to use them intuitively. Therefore site owners invest a lot of money on design and usability in order to get the most users. Currently usability testing usually is done in laboratories of user experience. During the interaction with web sites, data is collected (such as video record of screen or face of the user, clicks and mouse movements and oculography data). Evaluation of the results of such a study is difficult. Experts evaluate the collected data manually. For the purpose of reducing the effort of experts would be good to automate these evaluations as much as possible.

In our work, we focus on the automated identification of salient DOM elements, areas of interest and user-friendliness (detect the correct deployment of DOM elements, detect F-model) website. Solution has to be verified in the laboratory user experience and interaction of the faculty.

to the top | to the main

Automatic recognition of user´s characteristics based on eye-tracking data

Martina Redajová
master study, supervised by Jakub Šimko

Abstract. Design of user interface is complicated process requiring monitoring lot of factors. Since every user is different his behavior is often hard to predict. Despite of existence of certain ways of gathering information about users (registration forms), there are characteristics often impossible to detect, because they are changing duringtime period (emotions, skills, attention…).

In our work we focus on analysis of existing methods for recognition of users’ characteristics based on study of eye tracking data and use such method to automatically identify selected characteristics in user interface.

to the top | to the main

Identifying cognitive load

Lukáš Rešutík
bachelor study, supervised by Mária Bieliková

Abstract. New applications are created almost every day. If you want to work with these applications, it is good to know the user’s status by using them. For it has a great influence cognitive load.

In my bachelor thesis i am going to analyse methods of identification of cognitive load. One of the possible ways to identify cognitive load is carried out psychological test with a dual task. I will focus on the analysis of the types of secondary task, and then I will create an application that will be able to measure cognitive load, and verify it with other existing methods such as dilation of pupil.

Currently we are conducting experiments in UX Class with participants. Experiment consist of simple memory tasks.

to the top | to the main

HCI Literacy Estimation Using Eye Tracking

Monika Sanyová
master study, supervised by Eduard Kuric

Abstract. Nowadays testing quality of graphical user interface is very common. In interface quality testing it is useful to know about the user the most relevant information. One of the most interesting information about the user is the level of his HCI literacy estimated due to interaction with user interface and thanks to eye-tracking method we can retain detailed information about his interaction.

Our work focuses on HCI literacy estimation using eye tracking. The goal of this work is the design, development and experimental verification of classification method of user’s web HCI literacy estimation, which categorizes the user’s level of web HCl literacy into five levels based on his interaction with the components in our experimental graphical user interface (online shop prototype) during the execution of the tasks we defined. The goal is also to create a methodology for user’s web HCI literacy estimation, which makes our method applicable in other graphical user interfaces. Since baseline for our method validation is missing we also design a questionnaire for user’s web HCI literacy estimation.

to the top | to the main

Automatic Detection of Usability Smells Using Eye-Tracking

Martin Svrček
doctoral study, supervised by Mária Bieliková

Abstract. Nowadays, usability or user experience is very popular field and, especially in the context of web, there is a lot research papers in this area. Within this issue, the main purpose is to detect various anomalies of the design of applications and web pages. These anomalies are also called usability smells. Such detection is currently performed manually by experts on usability, who reveal these anomalies by evaluating the user studies.

The efforts of a number of studies in this field is to automate this process so such methods will be able to detect different usability smells. Nowadays, these methods and approaches used solely records of user activity in the form of clicks, scrolling or keyboard input. Our aim is to analyze a gaze of the users to determine the eye movement characteristics or metrics that can detect usability smells and improve the process of their identification. More specifically, we focused on websites.

to the top | to the main

User Experience and Emotions

Elena Štefancová
bachelor study, supervised by Róbert Móro

Abstract. User Experience is one of the most important criteria for designing user interfaces and testing their usability, with emotions as its essential element – how a user feels during interaction with an application. In our work, we measure their value and explore how they are influenced by noticeable mistakes of usability. In this work, two approaches are used: electroencephalography (EEG) and facial expression recognition. We use EEG device EMOTIV Epoc and for facial expression recognition we apply Noldus FaceReader software. We propose a method of classification of emotional changes from data obtained by this approach that employs also machine learning. To evaluate this method, we plan to carry out an experiment with an application, the interface of which includes intentionally inserted usability issues, and with which the users will work during the experiment. We expect this experiment to expose, what emotional response of the user could be caused by the usability issues.

to the top | to the main

Analysis of Reading Difficulty in Web Environment

Martin Štrbák
master study, supervised by Marián Šimko

Abstract. Success of Web applications depends on their ability to recognize user characteristics (interests, knowledge, type of personality etc.) and apply them to improve offered services. Mostly, these characteristics are acquired trough analysis of interaction of user with concrete Web application (where is clicking or looking, what is writing). More difficult is to determine if textual web content is interesting. One method to identify if reader is satisfied with content of document is to ask him. This approach has its limitations, because people are not always very honest, or have their own perception of classifying things as boring or entertaining. Questionnaire is a bit more objective. Most used technology nowadays is eye-tracker, which monitors human gaze during process of reading. Another methodology might be to analyze brain waves of user with electroencephalograph (EEG) while reading and compare them when his neural activity is idle. We will use this to determine whether text is difficult or easy to understand for human reader.

to the top | to the main