Ingeniero Civil Industrial / Este trabajo de título tiene por objetivo principal diseñar y desarrollar un módulo de clasificación de páginas web en base a las características de su contenido utilizando técnicas de minería de datos, lo que se traduce en la utilización de contenido HTML, análisis de texto visible de la página web y la incorporación de una variable que refleja la seguridad web según SSL como variables predictivas para la clasificación de páginas web. El trabajo se realiza enmarcado en el proyecto AKORI del Web Intelligence Centre de la Facultad de Ciencias Matemáticas de la Universidad de Chile, el cual pretende desarrollar una plataforma computacional para mejorar el diseño y contenido de sitios web mediante el estudio de variables fisiológicas y la aplicación de minería de datos. La plataforma consiste en la implementación de un modelo que sea capaz de predecir mapas tanto de fijación ocular como de dilatación pupilar de manera rápida y precisa. En esta etapa del proyecto AKORI es necesario mejorar el desempeño de las predicciones descritas, las cuales son realizadas en sitios web reales, de diseño y contenido muy variado. Además el comportamiento que se desea predecir es sobre usuarios de los que se desconoce su motivación para la navegación, lo cual a su vez altera tanto el comportamiento ocular como sus patrones de navegación. Dado lo anterior se propone como hipótesis de investigación: Es posible clasificar páginas web en base a las características de su contenido para solucionar dos problemas fundamentales, por un lado la clasificación agrupa páginas web maximizando la varianza de páginas web entre clases y minimizando la varianza intra clase, lo cual debiese mejorar considerablemente el desempeño del modelo, puesto que predecir dentro de una clase en la cual los ejemplos tienen mayor similitud disminuye el rango de error, disminuyendo, a su vez el error estándar en la predicción. Por otro lado entrega información sobre la motivación del usuario en la web si se conoce el servicio que ofrece la página web, lo que si bien no es información completa para describir el comportamiento del usuario, puede ser una importante variable de apoyo. Para el desarrollo del modelo se utiliza un juego de datos de 138 páginas web, escogidas según tráfico de usuarios Chilenos y luego se implementan cinco algoritmos de minería de datos para clasificar entre siete clases de páginas web. El algoritmo Naive Bayes obtiene el mejor desempeño, logrando un accuracy de 78.67%, lo que permite validar la hipótesis de investigación. Finalmente se concluye que se cumplen todos los resultados esperados y la hipótesis de investigación con resultados satisfactorios considerando la investigación actual.

Autism Spectrum Disorder (ASD) is often accompanied by atypical visual attention to faces. Previous studies have identified some predictors of atypical visual attention in ASD but very few have explored the role of conversational context. In this study, the fixation patterns of 19 typically developing (TD) children and 18 children with ASD were assessed during a SKYPED conversation where participants were asked to converse about mundane vs. emotion-laden topics. We hypothesized that 1) children with ASD would visually attend less to the eye region and more to the mouth region of the face compared to TD children and that 2) this effect would be exaggerated in the emotion-laden conversation. With regard to hypothesis 1, we found no difference between groups for either number of fixations or fixation time; however, children with ASD did evidence significantly more off-screen looking time compared to their TD peers. An additional analysis showed that compared to the TD group, the ASD group also had greater average fixation durations when looking at their speaking partner's face (both eyes and mouth) across conversational contexts. In support of hypothesis 2, eye tracking data (corrected for amount of time during conversation) revealed two interaction effects. Compared to the TD group, the ASD group showed 1) a decreased number of fixations to eyes and 2) an increased fixation time to mouths but only in the emotion-laden conversation. We also examined variables that predicted decreased number of eye fixations and increased mouth-looking in ASD in the emotion-laden conversation. Change scores (to be understood as the degree of visual attention shifting from the mundane to the emotion-laden condition) for the ASD group negatively correlated with age, perceptual reasoning skills, verbal ability, general IQ, theory of mind (ToM) competence, executive function (EF) subscales, and positively correlated with autism severity. Cognitive mechanisms at play and implications for theory and clinical practice are considered.

Individuals around the globe engage in sustainable consumption in their everyday life, e.g., when it comes to individual transportation. Although tourism behavior contributes to global carbon emissions to a considerable extent, consumers' awareness of sustainability in the tourism industry is still underresearched. Placing eco-labels next to tourist offers on websites might direct consumer's perception towards more sustainable offers. By employing eye-tracking techniques and surveys, this research aimed at linking information about sustainable tourist offers, perception of eco-labels and subsequent perception and preferences of tourism services. In Study 1, eight existing hotel offers with sustainability certification (four different labels) were selected and their websites presented to 48 participants (four websites each), whose eye movements were tracked. After looking at each website, they rated the overall appearance of the website. Based on the results, in the second study, participants' (n = 642) awareness of labels, their values and attitudes regarding sustainable behavior were found to influence their preference for certified tour operators. In addition, individuals' ideas of their perfect holidays were captured to allow a better understanding of their motivation. This research proposes implementing appropriate sustainable labeling in the tourism industry to increase awareness about sustainability among travelers and subsequently increase sustainable travel behavior.

Traditional theories of vision assume that object segregation occurs before access to object memories. Yet, behavioral evidence shows that familiar configuration is a prior for segregation, and electrophysiological experiments demonstrate these memories are accessed rapidly. A behavioral index of the speed of access is lacking, however. Here we asked how quickly behavior is influenced by object memories that are accessed in the course of object segregation. We investigated whether access to object memories on the groundside of a border can slow behavior during a rapid categorization task. Participants viewed two silhouettes that depicted a real-world and a novel object. Their task was to saccade toward the real-world object as quickly as possible. Half of the nontarget novel objects were ambiguous in that a portion of a real-world object was suggested, but not consciously perceived, on the groundside of their borders. The rest of the nontargets were unambiguous. We tested whether saccadic reaction times were perturbed by the real-world objects suggested on the groundside of ambiguous novel silhouettes. In Experiments 1 and 2, saccadic reaction times were slowed when nontargets were ambiguous rather than unambiguous. Experiment 2 set an upper limit of 190 ms on the time required for object memories in grounds to influence behavior. Experiment 3 ruled out factors that could have produced longer latencies other than access to object memories. These results provide the first behavioral index of how quickly memories of objects suggested in grounds can influence behavior, placing the upper limit at 190 ms.

Background: With the rise of streaming services such as cloud gaming, a fast internet speed is required for the overall experience. The average internet connection is not suited for the requirements that cloud gaming require. A high quality and frame rate is important for the experience. A solution to this problem would be to have parts where the user is looking at in a image be displayed in higher quality compared to the rest of the image. Objectives: The objective of this thesis is to create a gaze-based lossy image compression algorithm that reduces quality where the user is not looking. By using different radial functions to determine the quality decrease, the perceptual quality is compared to traditional JPEG compression. The storage difference when using a gaze-based lossy image compression is also compared to the JPEG algorithm. Methods: A gaze-based image compression algorithm, which is based on the JPEG algorithm, is developed with DirectX 12. The algorithm uses Tobii eye tracker to get where the user is gazing at the screen. When the gaze-position is changed the algorithm is run again to compress the image. A user study is conducted to the test the perceived quality of this algorithm compared to traditional lossy JPEG image compression. Two different radial functions are tested with various parameters to determine which one is offering the best perceived quality. The algorithm is also tested along with the radial functions on how much of a storage difference there is when using this algorithm compared to traditional JPEG compression. Results: With 11 participants, the results show the gaze-based algorithm is perceptually the same on images that have few objects who are close together. Images with many objects that are spread throughout the image performed worse on the gaze-based algorithm and was less picked compared traditional JPEG compression. The radial functions that cover much of the screen is more often picked compared to other radial functions that have less area of the screen. The storage difference between the gaze-based algorithm compared to traditional JPEG compression was between 60% to 80% less depending on the image. Conclusions: The thesis concludes that there is substantial storage savings that can be made when using a gaze-based image compression compared to traditional JPEG compression. Images with few objects who are close together are perceptually not distinguishable when using the gaze-based algorithm.

O trabalho pioneiro de Darwin, em 1872, constituiu uma das primeiras fontes de informação sobre as emoções dentro da perspectiva evolucionista. Hoje podemos contar com equipamentos de medidas mais sofisticados que nos permitem revisitar alguma das questões básicas por ele formuladas. A presente tese vem contribuir para uma melhor compreensão do reconhecimento de emoções e teoria da mente em crianças com TEA, apresentando analises que possam levar a sugestões de intervenções clínicas futuras. Embora ainda haja inconsistência entre os estudos quanto a qual emoção está mais prejudicada, tem sido relatadas dificuldades na identificação das emoções de medo, raiva, nojo e surpresa. Usamos um eye-tracker com o objetivo de esclarecer como os participantes olham para estímulos. Os participantes foram 40 meninos com TEA e com desenvolvimento típico, equiparados quanto a idade (6 a 14 anos). Foram avaliados para classificação do QI e para o grau de autismo. Os estímulos visuais foram retirados dos subteste de Reconhecimento de Emoções e Teoria da Mente do NEPSY II e apresentados em um equipamento de rastreamento de olhar, tendo sido definidas áreas de interesse (AI) do rosto. Para cada imagem, foi computado o tempo de fixação (em segundos) em cada AI. Os participantes também realizaram as tarefas do subteste, com o objetivo de verificar o acerto e o erro em cada estímulo. Em nossas análises pudemos verificar que a idade e o QI dos participantes influenciaram o desempenho na tarefa de reconhecimento de emoções e não influenciaram na tarefa de ToM. As crianças com TEA apresentaram um desempenho prejudicado quando comparado com as crianças com DT em ambos subtestes. Os participantes com TEA erraram mais as tarefas com emoções de valência negativa, como tristeza e raiva. A classificação do grau de TEA, não teve influência no desempenho dos participantes. O que parece ocorrer é que pertencer ao grupo clínico TEA seja uma condição para ter um desempenho pior quando comparado com o grupo de DT. Analisando o rastreamento de olhar, constatamos que o grupo DT permaneceu menos tempo fixando o olhar na AI dos olhos e mais tempo na AI da boca. Como o tempo total de fixação para os dois grupos não teve diferença estatística no subteste de reconhecimento de emoção, podemos sugerir que as crianças com TEA apresentam um padrão de olhar que chamamos de \"difuso\" enquanto que as crianças com DT apresentam um padrão de olhar que chamamos de \"focado\" nas regiões da face relevantes para o reconhecimento de emoções. Parece que na tarefa de ToM, o tempo de fixação em olhos e boca não foi o determinante para o erro na resposta e sim o olhar difuso na cena que deveria ser interpretada. Quando avaliamos os resultados qualitativos do rastreamento do olhar podemos pensar na importância de intervenções que possam ajudar crianças com TEA a melhorar seu desempenho em identificar as expressões faciais que auxiliam na decodificação da emoção e na interpretação de estados mentais / Darwin\'s pioneering work in 1872 was one of the earliest sources of information about emotions within the evolutionary perspective. Today we have more sophisticated equipment that allows us to revisit some of the basic questions that he formulates. The present thesis contributed to a better understanding of the emotion recognition and theory of mind in children with ASD, presenting analyzes that may lead to suggestions of future clinical interventions. Although there is still inconsistency among studies as to which emotion is most impaired, difficulties in identifying the emotions of fear, anger, disgust and surprise have been reported. We use an eye-tracker to clarify how participants look at stimuli. The participants were 40 boys with ASD and typical development, matched for age (6 to 14 years). They were evaluated for IQ and for the level of autism. The visual stimuli were from the Emotional Recognition and Theory of Mind subtests of NEPSY II and presented with an eye-tracking device, with areas of interest (AI) defined. For each image, the fixation duration (in seconds) in each AI was computed. The participants also performed the tasks of the subtest, in order to verify the right and wrong answers in each stimulus. In our analyzes we could verify that the participants\' age and IQ influenced the performance in the task of emotion recognition and did not influence the task of ToM. Children with ASD had an impaired performance when compared to children with TD in both subtests. Participants with ASD had a low score in tasks with negative valence emotions, such as sadness and anger. The classification of the ASD level had no influence on the participants\' performance. What seems to occur is that beeing on the ASD clinical group is a condition to had a worse perform when compared to the TD group. Analyzing the eye tracking, we found that the TD group spent less time fixing the eyes in the AI of the eyes and longer in the AI of the mouth. As the total fixation time for the two groups did not have a statistical difference in the emotion recognition subtest, we can suggest that children with ASD have a look pattern that we call \"diffuse\", while children with TD have a look pattern which we call \"focused\" on regions of the face relevant to the recognition of emotions. It seems that in the ToM task, the fixation time in the eyes and mouth was not the determinant for the error in the response, but the diffuse look in the scene that was to be interpreted. When we evaluate the qualitative results of eye tracking we can think of the importance of interventions that can help children with ASD improve their performance in identifying the facial expressions that aid in the decoding of emotion and the interpretation of mental states

Gaze estimation systems determine where someone is looking. Gaze is used for a wide range of applications including market research, usability studies, and gaze-based interfaces. Traditional equipment uses special hardware. To bring gaze estimation mainstream, researchers are exploring approaches that use commodity hardware alone. My work addresses two outstanding problems in this field: 1) it is hard to collect good ground truth eye images for machine learning, and 2) gaze estimation systems do not generalize well -- once they are trained with images from one scenario, they do not work in another scenario. In this dissertation I address these problems in two different ways: learning-by-synthesis and analysis-by-synthesis. Learning-by-synthesis is the process of training a machine learning system with synthetic data, i.e. data that has been rendered with graphics rather than collected by hand. Analysis-by-synthesis is a computer vision strategy that couples a generative model of image formation (synthesis) with a perceptive model of scene comparison (analysis). The goal is to synthesize an image that best matches an observed image. In this dissertation I present three main contributions. First, I present a new method for training gaze estimation systems that use machine learning: learning-by-synthesis using 3D head scans and photorealistic rendering. Second, I present a new morphable model of the eye region. I show how this model can be used to generate large amounts of varied data for learning-by-synthesis. Third, I present a new method for gaze estimation: analysis-by-synthesis. I demonstrate how analysis-by-synthesis can generalize to different scenarios, estimating gaze in a device- and person- independent manner.

Many models of reading include cognitive components such as memory and attention in addition to the linguistic processing aspects; however, the particular effects of these other resources have yet to be clearly defined. The current study seeks to look at the effects of attentional resource manipulation on the reading process through the use of comic viewing, auditory distraction and dual-task paradigms. The study uses both behavioral measures (comprehension question performance) and eye-tracking measures (fixation time, revisits) to investigate the effects of increased cognitive load on the reading process in healthy, literate adults. The study found that healthy adults used the presence of comics to attenuate the effects of distraction and increased cognitive load on the comprehension process. Without the additional visual information of the comics, the participants showed a significant increase in the number of attempts to reread material (i.e. revisits) due to the distractor/dual-task presentation while still exhibiting notable decreases in comprehension. Furthermore, additional correlations were found between several cognitive testing measures and the participants’ performances in both the comprehension and eye-tracking measures across conditions. The results of this study provide valuable insight into future stages of this study, informing further research with both healthy adults and clinical populations such as people with aphasia or traumatic brain injuries.

A concussion can result in a wide range of signs and symptoms including physical, cognitive, emotional and sleep-related symptoms. While a concussion can occur at any age, children and adolescents are at an increased risk of experiencing a concussion. Important physical consequences of concussion are balance deficits which affect approximately one in three cases. Balance is essential to perform many postures and activities and therefore it is important to accurately identify these deficits at the time of injury and during the recovery process in order to avoid potential further injury. A review of the literature regarding the different clinical and laboratory measures used to assess balance following a concussion was completed and identified several gaps in the literature. The four studies in this thesis were tailored to address these gaps in knowledge. The first study addressed the sensitivity of a gold standard clinical measure and laboratory measures of standing balance in terms of identifying balance deficits in adolescents at 1-month post-concussion. Performance on the Balance Error Scoring System (BESS) and COP measures from single and dual-task balance conditions were compared between a group of adolescents at one-month post-concussion and a group of non-injured adolescents. The results demonstrated that COP measures from single and dual-task conditions identified balance deficits in the concussed group that were not captured by the BESS. The second study addressed whether self-reported balance problems are a suitable alternative to COP measures to identify balance deficits in concussed adolescents. A secondary analysis of the data from study 1 was conducted to compare COP measures from the single and dual-task conditions between concussed adolescents self-reporting balance problems, concussed adolescents self-reporting no balance problems and a group of non-injured adolescents. The results from this analysis showed that the concussed adolescents demonstrated balance deficits regardless of whether they self-reported balance problems. There is significant clinical interest that lies in the ability to predict which adolescents presenting in the emergency department with concussion will be affected with ongoing balance deficits. The third study addressed the predictive ability of a set of COP variables recorded within the first 10 days following injury in a group of concussed adolescents to predict balance performance on a dual-task condition at one-month post-injury. Seven COP variables were identified as significant predictors. A secondary objective of this study was to compare performance on the COP measures between the concussed adolescents and a group of non-injured adolescents during the first session within the first 10 days following injury and during the second session at one-month post-injury and to compare performance between sessions within each group. The between session comparisons showed that performance remained relatively stable across sessions within both groups. In contrast, the between group comparisons revealed several significant differences in COP measures between the concussed and non-injured groups. The fourth study in this thesis addressed the association between balance and saccadic eye movements in concussed adolescents. Impaired saccadic eye movements are an important consequence of concussion and may be associated with balance deficits since both processes are dependent on several of the same cortical structures and brainstem areas. In this study, concussed and non-injured adolescents performed three different dual-task balance conditions involving either a high cognitive load, a low cognitive load and a gaze shifting component or a high cognitive load and a gaze shifting component. The results demonstrated that the concussed adolescents swayed over larger 95% ellipse areas while performing the two dual-tasks with the gaze shifting component, but these larger amounts of sway were not associated with an increase in saccades. Taken together, these four studies extend the current knowledge regarding balance performance in concussed adolescents and provide results that can be applied to balance assessments for concussion.

Web pages are typically comprised of different kinds of visual elements such as menus, headers and footers. To improve user experience, eye tracking has been widely used to investigate how users interact with such elements. In particular, eye movement sequences, called scanpaths, have been analysed to understand the path that people follow in terms of these elements. However, individual scanpaths are typically complicated and they are related to specific users, and therefore any processing done with those scanpaths will be specific to individuals and will not be representative of multiple users. Therefore, those scanpaths should be clustered to provide a general direction followed by users. This direction will allow researchers to better understand user interactions with web pages, and then improve the design of the pages accordingly. Existing research tends to provide a very short scanpath which is not representative for understanding user behaviours. This thesis introduces a new algorithm for clustering scanpaths, called Scanpath Trend Analysis (STA). In contrast to existing research, in STA, if a particular element is not shared by all users but it gets at least the same attention as the fully shared elements, it is included in the resulting scanpath. Thus, this algorithm provides a richer understanding of how users interact with web pages. The STA algorithm was evaluated with a series of eye tracking studies where the web pages used were automatically segmented into their visual elements by using different approaches. The results show that the outputs of the STA algorithm are significantly more similar to the inputted scanpaths in comparison with the outputs of other existing work, and this is not limited to a particular segmentation approach. The effects of the number of users were also investigated on the STA algorithm as the number of users required for scanpath analysis has not been studied in depth in the literature. The results show the possibility to reach the same results with a smaller group of users. The research presented in this thesis should be of value to eye tracking researchers, to whom the STA algorithm has been made available to analyse scanpaths, and to behaviour analysis researchers, who can use the algorithm to understand user behaviours on web pages, and then design, develop and present the pages accordingly.