Making a robot autonomous has been a common challenge to be overcome since the very beginning. To be an autonomous system, the robot should collect environmental data, interpret them, and act accordingly. In order to accomplish these, some resource management should be conducted. That is, the resources, which are time, and computation power in our case, should be allocated to more important areas. Existing researches and approaches, however, are not always human like. Indeed they don&rsquo / t give enough importance on this. Starting from this point of view, the system proposed in this thesis supplies the resource management trying to be more &rsquo / human like&rsquo / . It directs the focus of attention to where higher resolution algorithms are really needed. This &rsquo / real need&rsquo / is determined by the visual features of the scene, and current importance levels (or weight values) of each of these features. As a further attempt, the proposed system is compared with human subjects&rsquo / characteristics. With unbiased subjects, a set of parameters which resembles a normal human is obtained. Then, in order to see the effect of the guidance, the subjects are asked to concentrate on a single predetermined feature. Finally, an artificial neural network based learning mechanism is added to learn to mimic a single human or a group of humans. The system can be used as a preattentive stage module, or some more feature channels can be introduced for better performance in the future.

We present a post-processing technique that selectively reduces the salience of distracting regions in an image. Computational models of attention predict that texture variation influences bottom-up attention mechanisms. Our method reduces the spatial variation of texture using power maps, high-order features describing local frequency content in an image. Modification of power maps results in effective regional de-emphasis. We validate our results quantitatively via a human subject search experiment and qualitatively with eye tracking data. / Singapore-MIT Alliance (SMA)

A critical function of the human visual system is to parse objects from the larger context of the environment, allowing for the identification of, and potential interaction with, those objects. The use of contextual information allows us to rapidly locate, identify, and interact with objects that appear in the environment. Contextual information can help specify an object's location within the environment (allocentric encoding) or with respect to the observer (egocentric encoding). Understanding how contextual information influences perceptual organization, and the neural systems that process a complex scene, is critical in understanding how contextual information assists in parsing local information from background. In the real world, relying on context is typically beneficial, as most objects occur in circumscribed environments. However, there are circumstances in which context can harm performance. In the case of visual illusions, relying on the context can bias observers' perceptions and cause significant motor errors. Studying the illusory conditions under which perceptual/motor functions are "fooled", or breakdown, can provide valuable information about how the brain computes allocentric and egocentric frames of reference. The following studies examine how attentional (Chapters II & III) manipulations of visuospatial context affect components of observers' egocentric reference frames (e.g., perceived vertical or subjective midline) and how neural manipulations (Chapter IV) can modulate observers' reliance on contextual information. In Chapter II, the role of attentional control settings on contextual processing is examined. Chapter III addresses the question of how visuospatial shifts of attention interact with an egocentric frame of reference. Finally, Chapter IV examines the functional role of superior parietal cortex in the processing of egocentric contextual information.

Les humains reçoivent grande quantité d'informations de l'environnement avec vue et l'ouïe . Pour nous aider à réagir rapidement et correctement, il existe des mécanismes dans le cerveau à l'attention de polarisation vers des régions particulières , à savoir les régions saillants . Ce biais attentionnel n'est pas seulement influencée par la vision , mais aussi influencée par l'interaction audio - visuelle . Selon la littérature existante , l'attention visuelle peut être étudié à mouvements oculaires , mais l'effet sonore sur le mouvement des yeux dans les vidéos est peu connue . L'objectif de cette thèse est d'étudier l'influence du son dans les vidéos sur le mouvement des yeux et de proposer un modèle de saillance audio - visuel pour prédire les régions saillants dans les vidéos avec plus de précision . A cet effet, nous avons conçu une première expérience audio - visuelle de poursuite oculaire . Nous avons créé une base de données d'extraits vidéo courts choisis dans divers films . Ces extraits ont été consultés par les participants , soit avec leur bande originale (condition AV ) , ou sans bande sonore ( état ​​V) . Nous avons analysé la différence de positions de l'oeil entre les participants des conditions de AV et V . Les résultats montrent qu'il n'existe un effet du bruit sur le mouvement des yeux et l'effet est plus important pour la classe de la parole à l'écran . Ensuite , nous avons conçu une deuxième expérience audiovisuelle avec treize classes de sons. En comparant la différence de positions de l'oeil entre les participants des conditions de AV et V , nous concluons que l'effet du son est différente selon le type de son , et les classes avec la voix humaine ( c'est à dire les classes parole , chanteur , bruit humain et chanteurs ) ont le plus grand effet . Plus précisément , la source sonore a attiré considérablement la position des yeux uniquement lorsque le son a été la voix humaine . En outre , les participants atteints de la maladie de AV avaient une durée moyenne plus courte de fixation que de l'état de V . Enfin , nous avons proposé un modèle de saillance audio- visuel préliminaire sur la base des résultats des expériences ci-dessus . Dans ce modèle , deux stratégies de fusion de l'information audio et visuelle ont été décrits: l'un pour la classe de son discours , et l'autre pour la musique classe de son instrument . Les stratégies de fusion audio - visuelle définies dans le modèle améliore la prévisibilité à la condition AV / Humans receive large quantity of information from the environment with sight and hearing. To help us to react rapidly and properly, there exist mechanisms in the brain to bias attention towards particular regions, namely the salient regions. This attentional bias is not only influenced by vision, but also influenced by audio-visual interaction. According to existing literature, the visual attention can be studied towards eye movements, however the sound effect on eye movement in videos is little known. The aim of this thesis is to investigate the influence of sound in videos on eye movement and to propose an audio-visual saliency model to predict salient regions in videos more accurately. For this purpose, we designed a first audio-visual experiment of eye tracking. We created a database of short video excerpts selected from various films. These excerpts were viewed by participants either with their original soundtrack (AV condition), or without soundtrack (V condition). We analyzed the difference of eye positions between participants with AV and V conditions. The results show that there does exist an effect of sound on eye movement and the effect is greater for the on-screen speech class. Then, we designed a second audio-visual experiment with thirteen classes of sound. Through comparing the difference of eye positions between participants with AV and V conditions, we conclude that the effect of sound is different depending on the type of sound, and the classes with human voice (i.e. speech, singer, human noise and singers classes) have the greatest effect. More precisely, sound source significantly attracted eye position only when the sound was human voice. Moreover, participants with AV condition had a shorter average duration of fixation than with V condition. Finally, we proposed a preliminary audio-visual saliency model based on the findings of the above experiments. In this model, two fusion strategies of audio and visual information were described: one for speech sound class, and one for musical instrument sound class. The audio-visual fusion strategies defined in the model improves its predictability with AV condition.

This thesis is all about the visual attention, starting from understanding the human visual system up till applying this mechanism to a real-world computer vision application. This has been achieved by taking the advantage of latest findings about the human visual attention and the increased performance of the computers. These two facts played a vital role in simulating the many different aspects of this visual behavior. In addition, the concept of bio-inspired visual attention systems have become applicable due to the emergence of different interdisciplinary approaches to vision which leads to a beneficial interaction between the scientists related to different fields. The problems of high complexities in computer vision lead to consider the visual attention paradigm to become a part of real time computer vision solutions which have increasing demand.  In this thesis work, different aspects of visual attention paradigm have been dealt ranging from the biological modeling to the real-world computer vision tasks implementation based on this visual behavior. The implementation of traffic signs detection and recognition system benefited from this mechanism is the central part of this thesis work.

Book Summary: The international nature of the sport industry presents many management challenges and opportunities for sport organisations. This book brings together cutting-edge research from leading sport management scholars around the world, surveying a wide range of topics and issues facing the sport industry today. It represents an essential platform for the international exchange of ideas, best practice and research in sport management studies. The globalisation of the sport industry has brought increased complexity to organisations’ operations in terms of regulation, competition and multiculturalism. Drawing on a wealth of original research from fifteen countries, this book addresses a variety of global, regional, national and community issues that are central to successful sport management. Combining both qualitative and quantitative studies, it explores key themes, such as managing resources and organisational change, marketing and promotion, law and regulation, sport-for-development and research protocols. Global Sport Management Studies: Contemporary issues and inquiries is essential reading for all students and scholars of sport management, sport business and sport marketing, as well as for any professional working in the sport and leisure industries.

The ability to visually explore the world is a substantial and crucial factor for humans to interact and navigate in their environment. Given the importance of the visual system, it is no wonder that vision research has become one of the major fields in cognitive science. Eye tracking studies hereby revealed three main factors, that guide our visual perception and attention. These factors can be described as bottom-up and top-down features, as well as spatial properties. Whereas previous research has put a lot of effort in investigating each of these factors in detail, the combined interaction of the factors - especially regarding spatial properties and biases - is less well understood. This thesis deals with the different aspects of several features influencing visual behavior. I present four studies that examine different aspects of images with respect to their spatial properties. Hereby, I elaborate the global salience of images as spatial factor, which can be impacted by top-down features, such as emotions. Further, I describe how the effect of image sizes as a spatial property is linked with visual exploration and exploitation. In study 1 (Chapter 2) we investigated the global salience of images and its properties. While recording of eye movements, participants freely observed several image pairs, where one image has been shown on the left, and one image on the right side of the screen. Based on results of the eye tracking recordings, we trained a logistic regression model to calculate a global salience coefficient for each image that can be ranked in order to predict the location - left or right image - of the first fixations in an image pair. Our trained model was able to accurately predict the first fixation of an image pair, indicating that images indeed provide a unique global salience score. Hereby, we showed that the global salience of an image is independent of local salience and further that a given task or familiarity of the images affected the respective global salience. In study 2 (Chapter 3) we investigated the influence of emotions as top-down factor on natural visual behavior. Participants had been emotionally primed either by a sequence of positive or negative laden images. Afterwards, we recorded participants' eye movements while being confronted with image pairs, were one positive and one negative image - left and right respectively - was presented simultaneously. Results showed that positively primed participants tended to shift overt attention towards negative images in the stimulus pairs. In a later memory request, we further showed that such participants had better recall performance of negative image content. Thus, we concluded that positive priming increases attention and memory towards negative content. In study 3 (Chapter 4) we investigated the trade-off between the exploration (i.e. initiate fixations to unseen image areas) and exploitation (i.e. stay and process the currently fixated information in depth) of natural images with varying sizes. Participants freely observed images of different sizes and categories while eye movements have been recorded. For exploration we tested the distribution of fixations measured by the central tendency and entropy, as well as number of fixations and saccade amplitudes. The exploitation was derivated from the fixation duration. In our results, we found that larger image sizes led to a shift from exploitation towards exploration. That is, images have been explored more broaden while in-depth processing reduced accordingly. In study 4 (Chapter 5) we investigated the effect of image sizes in natural viewing behavior within patients suffering from retinitis pigmentosa (RP); an inherited disease that causes progressive peripheral visual-field loss. The purpose of this study was to examine whether visual behavior differed when the scene content was shown in various extents of the perceived visual field. For this, participants with varying degrees of visual-field loss and healthy control participants freely observed images of different sizes while eye movements were recorded. For healthy control participants we could replicate the effects already described in Chapter 4. That is, larger images lead to a shift from exploitation to exploration. Surprisingly, on group level RP patients scanned the images similar to the healthy control participants. However, on individual level RP patients also showed individual idiosyncratic explorative strategies when the observed scene exceeded their visible field. We thus concluded, that although retinitis pigmentosa leads to a severe loss of the visual field, there seems to be no general adaptive mechanism adapting visual exploration accordingly. Instead, individuals rely on individual strategies, leading to high heterogeneity in the RP group. In a nutshell, our results showed that images provide unique global salience coefficients that can predict attraction of visual attention. However, emotions as top-down factor affect these global salience coefficients as shown by a shift of attention towards negative image content. The size of an image as spatial property affects natural viewing behavior in such way that in-depth processing (exploitation) shifts towards broader exploration in large images. These results remain remarkably stable, even when the image size extends the visual field (in RP patients). Summarized, the studies in this thesis were motivated to gain and deepen knowledge in the interplay between distinct factors - with an emphasis on spatial properties - that impact visual behavior and attention.

京都大学 / 0048 / 新制・課程博士 / 博士(人間・環境学) / 甲第19079号 / 人博第732号 / 新制||人||176(附属図書館) / 26||人博||732(吉田南総合図書館) / 32030 / 京都大学大学院人間・環境学研究科共生人間学専攻 / (主査)教授 齋木 潤, 教授 船橋 新太郎, 准教授 月浦 崇 / 学位規則第4条第1項該当 / Doctor of Human and Environmental Studies / Kyoto University / DGAM

In the recent years, a huge success has been accomplished in prediction of human eye fixations. Several studies employed deep learning to achieve high accuracy of prediction of human eye fixations. These studies rely on pre-trained deep learning for object classification. They exploit deep learning either as a transfer-learning problem, or the weights of the pre-trained network as the initialization to learn a saliency model. The utilization of such pre-trained neural networks is due to the relatively small datasets of human fixations available to train a deep learning model. Another relatively less prioritized problem is amount of computation of such deep learning models requires expensive hardware. In this dissertation, two approaches are proposed to tackle abovementioned problems. The first approach, codenamed DeepFeat, incorporates the deep features of convolutional neural networks pre-trained for object and scene classifications. This approach is the first approach that uses deep features without further learning. Performance of the DeepFeat model is extensively evaluated over a variety of datasets using a variety of implementations. The second approach is a deep learning saliency model, codenamed ClassNet. Two main differences separate the ClassNet from other deep learning saliency models. The ClassNet model is the only deep learning saliency model that learns its weights from scratch. In addition, the ClassNet saliency model treats prediction of human fixation as a classification problem, while other deep learning saliency models treat the human fixation prediction as a regression problem or as a classification of a regression problem.

The research described here examines infant and parent attention to a familiar baby video. Also of interest, was if infant viewing behaviors influenced parent viewing behaviors, and vice versa. Subjects were 12-15 and 18-21 month-old infants who were observed watching a familiar baby video with one parent. Overall infants and adults spent less than one-third of the time watching the television. This measure varied greatly across dyads. However, there was a strong, positive relationship within dyads, suggesting that infants and parents may be influencing each other’s viewing behavior. Further analyses revealed that there was a social component that influenced when infants and parents initiated and terminated looks to the television that extended above and beyond the common influence of the formal features of the program. Though this influence was mutual for both the infant and parent, overall, infants tended to ‘lead’ their parents’ looks more frequently than parents’ led their infants’.