Dissertations / Theses on the topic 'Page layout'

Knowledge elicitation remains a fundamental feature of Knowledge Based Systems evolution. However, there is insufficient evidence to support the presumption that the knowledge elicitation philosophy is viable for design. Scant effort has been applied to research into techniques for design elicitation, and the nature of design is poorly considered. In particular, design tasks that involve visual design skills appear especially neglected. The scarcity of proven knowledge elicitation methods for design has not dampened the enthusiasm for "Intelligent" Computer Aided Design Systems. However, it is argued that design knowledge acquired from ad hoc, unsubstantiated and untested procedures, and knowledge that is undocumented and untested cannot be considered reliable. Indeed, it is extensively observed that a deficiency of intelligent performance exists in current ICAD systems, and the exigency for laudable design elicitation methods is prevalent Here, knowledge elicitation in design is promoted through a review of design and knowledge elicitation research literature. Design must be considered dissimilar to scientific problem solving, and the holistic nature of the task is an important characteristic. Further, the spatial, diagrammatical and drawing forms of communication, that are manifest in design, must be tackled. A method for the elicitation of design knowledge is proposed, and tested in the domain of page layout design. Computerised methods of knowledge acquisition currently lack the sophistication to expound the enigmas associated with design elicitation. It is concluded that the personal interview strategy is appropriate, in which the nature of the design task, and the visual and spatial components are equitably considered. The understanding of page layout design is demonstrated in a communicable report, and tested through an evaluation study. It is concluded that methodological principles of knowledge elicitation are appropriate to design, and a suitable method is outlined. The domain of page layout design illustrates that the techniques are successful, useful and practical.

This thesis was a descriptive study of the process of designing a newspaper front page by three journalists. The newspaper industry has placed increased importance on journalistic design. However, little research had been completed on the mental processes of designers.This study was concerned with the processes that occur as professional journalists designed a front page of a newspaper as determined by protocol analysis. Three Indiana newspapers,The Evansville Courier, The South Bend Tribune and The Times from Munster, Ind., were selected. The individual who designed the front page on a regular basis participated in the protocol sessions.The conclusions showed that these designers spent the majority of the design process planning and analyzing materials to be used. The majority of all decisions made by the designers were based on past design knowledge. The journalists applied their skills in a rapid and confident way as they designed separate portions of the page until all elements were used and the page was complete.
Department of Journalism

In this thesis, we consider three parameters associated with graphs : stack number, track number, and layered pathwidth. Our first result is to show that the stack number of any graph is at most 4 times its layered pathwidth. This result complements an existing result of Dujmovic et al. that showed that the queue number of a graph is at most 3 times its layered pathwidth minus one (Dujmovic, Morin, and Wood [SIAM J. Comput., 553–579, 2005]). Our second result is to show that graphs of track number at most 3 have layered pathwidth at most 4. This answers an open question posed by Banister et al. (Bannister, Devanny, Dujmovic, Eppstein, and Wood [GD 2016, 499–510, 2016, Algorithmica, 1–23, 2018]).

The aim was to explore how the first page of a student newspaper has been changed over time since it was founded in the 1960´s. This thesis explores how the front page of the student newspaper Lösnummer has been changed since it was founded in 1966. It concerns the location and frequency of elements and it is based on theories of layout, the first page and visual communication In addition, the thesis develops a new kind of method for determining the location of elements on the type area. The method is based on the hexadecimal number system. It is formed as a grid which is placed on the first page to determine the location of the elements. It concludes that the frequency of elements had its peak during the 1970´s and has ever since decreased gradually. Image placement has gone from being scattered all over the front page to be centered and increase in size.
Syftet med uppsatsen var att undersöka hur en studenttidning har förändrats sedan den grundades på 1960-talet. Denna avhandling undersöker hur framsidan av studenttidningen Lösnummer har förändrats sedan den grundades 1966. Den behandlar placering och frekvens av element och är baserad på teorier om layout, första sidan och visuell kommunikation. Avhandlingen utvecklar en ny typ av metod för fastställande av placeringen av elementen på förstasidan. Metoden bygger på det hexadecimala talsystemet då ett rutnät placeras på första sidan för att bestämma placeringen av elementen. Slutsatsen av undersökningen visare att frekvensen av elementen hade sin höjdpunkt under 1970-talet och har sedan minskat successivt. Bildernas placering har gått från att vara utspridda över hela framsidan till vara centreras och öka i storlek.

Background. This thesis focuses on the task of historical document semantic segmentation with recurrent neural networks. Document semantic segmentation involves the segmentation of a page into different meaningful regions and is an important prerequisite step of automated document analysis and digitisation with optical character recognition. At the time of writing, convolutional neural network based solutions are the state-of-the-art for analyzing document images while the use of recurrent neural networks in document semantic segmentation has not yet been studied. Considering the nature of a recurrent neural network and the recent success of recurrent neural networks in document image binarization, it should be possible to employ a recurrent neural network for document semantic segmentation and further achieve high performance results. Objectives. The main objective of this thesis is to investigate if recurrent neural networks are a viable alternative to convolutional neural networks in document semantic segmentation. By using a combination of a convolutional neural network and a recurrent neural network, another objective is also to determine if the performance of the combination can improve upon the existing case of only using the recurrent neural network. Methods. To investigate the impact of recurrent neural networks in document semantic segmentation, three different recurrent neural network architectures are implemented and trained while their performance are further evaluated with Intersection over Union. Afterwards their segmentation result are compared to a convolutional neural network. By performing pre-processing on training images and multi-class labeling, prediction images are ultimately produced by the employed models. Results. The results from the gathered performance data shows a 2.7% performance difference between the best recurrent neural network model and the convolutional neural network. Notably, it can be observed that this recurrent neural network model has a more consistent performance than the convolutional neural network but comparable performance results overall. For the other recurrent neural network architectures lower performance results are observed which is connected to the complexity of these models. Furthermore, by analyzing the performance results of a model using a combination of a convolutional neural network and a recurrent neural network, it can be noticed that the combination performs significantly better with a 4.9% performance increase compared to the case with only using the recurrent neural network. Conclusions. This thesis concludes that recurrent neural networks are likely a viable alternative to convolutional neural networks in document semantic segmentation but that further investigation is required. Furthermore, by combining a convolutional neural network with a recurrent neural network it is concluded that the performance of a recurrent neural network model is significantly increased.
Bakgrund. Detta arbete handlar om semantisk segmentering av historiska dokument med recurrent neural network. Semantisk segmentering av dokument inbegriper att dela in ett dokument i olika regioner, något som är viktigt för att i efterhand kunna utföra automatisk dokument analys och digitalisering med optisk teckenläsning. Vidare är convolutional neural network det främsta alternativet för bearbetning av dokument bilder medan recurrent neural network aldrig har använts för semantisk segmentering av dokument. Detta är intressant eftersom om vi tar hänsyn till hur ett recurrent neural network fungerar och att recurrent neural network har uppnått mycket bra resultat inom binär bearbetning av dokument, borde det likväl vara möjligt att använda ett recurrent neural network för semantisk segmentering av dokument och även här uppnå bra resultat. Syfte. Syftet med arbetet är att undersöka om ett recurrent neural network kan uppnå ett likvärdigt resultat jämfört med ett convolutional neural network för semantisk segmentering av dokument. Vidare är syftet även att undersöka om en kombination av ett convolutional neural network och ett recurrent neural network kan ge ett bättre resultat än att bara endast använda ett recurrent neural network. Metod. För att kunna avgöra om ett recurrent neural network är ett lämpligt alternativ för semantisk segmentering av dokument utvärderas prestanda resultatet för tre olika modeller av recurrent neural network. Därefter jämförs dessa resultat med prestanda resultatet för ett convolutional neural network. Vidare utförs förbehandling av bilder och multi klassificering för att modellerna i slutändan ska kunna producera mätbara resultat av uppskattnings bilder. Resultat. Genom att utvärdera prestanda resultaten för modellerna kan vi i en jämförelse med den bästa modellen och ett convolutional neural network uppmäta en prestanda skillnad på 2.7%. Noterbart i det här fallet är att den bästa modellen uppvisar en jämnare fördelning av prestanda. För de två modellerna som uppvisade en lägre prestanda kan slutsatsen dras att deras utfall beror på en lägre modell komplexitet. Vidare vid en jämförelse av dessa två modeller, där den ena har en kombination av ett convolutional neural network och ett recurrent neural network medan den andra endast har ett recurrent neural network uppmäts en prestanda skillnad på 4.9%. Slutsatser. Resultatet antyder att ett recurrent neural network förmodligen är ett lämpligt alternativ till ett convolutional neural network för semantisk segmentering av dokument. Vidare dras slutsatsen att en kombination av de båda varianterna bidrar till ett bättre prestanda resultat.

Les graphistes utilisent des propriétés visuelles comme la couleur, la police de caractères typographiques, la taille ou la position pour construire et communiquer leurs concepts visuels. La plupart des outils numériques de composition considèrent les propriétés visuelles comme simples attributs. Ces propriétés sont accessibles, quand l’objet est sélectionné,via des boîtes de dialogue, sans êtres considérées comme objets indépendants dans l’interface. Cette thèse porte sur la réification des propriétés visuelles des objets graphiques pour les tâches de composition.Je soutiens l’importance de créer des outils de composition qui permettent de manipuler les propriétés visuelles directement, sans passer par des menus ou des boîtes de dialogue.La manipulation des propriétés visuelles dépend de la perception des propriétés, de l’expertise de l’utilisateur et de leur représentations sur l’interface.Elle dépend également des outils existants pour manipuler les propriétés visuelles dans le contexte de travail. Les outils numériques de composition utilisent principalement des boîtes de dialogue. Certains travaux de recherche soutiennent l’importance de faire évoluer les boîtes de dialogues, et proposent des boîtes de dialogues plus interactives. D’autres travaux proposent de nouvelles manières de représenter et de manipuler les propriétés visuelles.Les graphistes préfèrent les outils qui agissent directement sur les propriétés visuelles. Ils adaptent ces propriétés, pour les utiliser comme outils leur permettant de compléter des tâches comme l’alignement et l’espacement.Les artistes et les designers travaillent en développant leurs propres représentations des espaces de couleur. Ils manipulent chaque couleur, dans le contexte des autres couleurs utilisées dans le même document.Leurs choix de couleur dépendent également d’autres propriétés visuelles comme la texture, la matière, ou la luminosité. En créant leurs mise en page, les graphistes établissent des liens entre les propriétés spatiales comme la position, la taille, et la superposition des éléments graphiques. Ils créent des règles pour définir comment ces liens évoluent dans un même document, ou entre différents documents.Deux groupes d’outils proposent des exemples de la réification des propriétés visuelles pour des tâches de composition. Deux études avec des designers présentent différentes utilisations envisagées pour ces outils.Chaque utilisateur présente des exemples de projets spécifiques dans lesquels ces outils faciliteront leurs tâches de composition, et comment ils leur permettront d’obtenir les rendus recherchés
In this dissertation, I argue that visual composition tools should reify visual properties— that is, create them into first-class, interactive objects that designers can manipulate, directly in the document. Artists and designers use visual properties such as color, typography,size, and position to create novel composition concepts. Most visual composition tools treat graphical elements as objects— but not their visual properties. The latter are simply attributes of graphical elements, usually accessible on demand, through property sheets or dialog boxes, rather than as independen tinteractive objects.I begin with an introduction, where I outline the dissertation and summarise the main contributions.Then, I present a conceptual overview, where I argue that perception, expertise, theories and existing representations of visual properties affect how designers manipulate them in their work. In the related worksection, I identify the main approaches visual composition tools propose to manipulate visual properties.I explain the advantages and limitations of these approaches.I present a comparative structured observation study (Chapter 4) where graphic designers perform visual composition tasks in Adobe Illustrator. Designers preferred tools that provide direct access to visual properties.They use these visual properties to complete visual composition tasks such as alignment and distribution.I then present two interview studies (Chapter5, Chapter 6) that examined artists’ and designers’practices as they manipulate color and create layouts in their projects. Artists and designers create personal color representations. They manipulate each color in the context of its surrounding graphical elements,and combine it with other visual properties such as texture. As they create their layouts, designers establish links among visual properties such as size,position, and layering of graphical elements. They define rules for how these properties change in space,across instances of the same composition, or in time,across related compositions.Based on these observations, two groups of prototypes (Chapter 7) demonstrate how we can reify visual properties into first-class graphical objects. ColorPartner supports creating personal representations of the color manipulation process. Color Revealer permits creating informative color representations of how a process evolves. Palette Explorer supports manipulatingvisual properties of color swatches, such as size, position and layering, in the context of surrounding colors. Color Compositor supports manipulating the combination of colors and textures. Linkify supports linking visual properties directly in the documentand defining rules for how these links evolve as the content changes. Contextify supports creating rules that define how visual properties change across instances of a composition. In two follow up studies,designers explain how they intend to use these tools in their current projects..I conclude with a summary of the contributions(Chapter 8) and a discussions of the limitations and future work

Aim: The purpose of this study is to identify factors that can attract customers to buy online and help the grocers to improve their online grocery services to retain and obtain customers. Design/Methodology/Approach: A qualitative method was used, based on triangulation methodology including observation and interviews of manager and employees. Quantitative data was collected by handing out questionnaire to 204 potential customers and 30 existing customers. Findings: The study shows that product and service quality, time saving and convenience, web page layout, customers’ trust in grocers (service providers), store image and advertising are the important factors that customers consider when they decide to use online grocer service. E- grocers also need to consider these factors when they start online grocery service. Suggestions for Future Research: An extensive study with large population sample and co- operation from other companies can help to generalize the results. Future studies can be extended to include drive-in and shared reception box service method, and consider the cost for the customers in e-grocery sector.

This thesis deals with the design and implementation of the page layout manager for a web user interface composed of components. A part of solution is also a new XML format for configuring of the manager of page layout and composition and a new application interface for components. The created system has met the requirements of the European project KiWi and has passed the acceptance tests of its project team.

Submitted by Maike Costa (maiksebas@gmail.com) on 2016-01-06T14:43:30Z No. of bitstreams: 1 arquivototal.pdf: 1912842 bytes, checksum: dbfd74608eda29f9015d9d5015514d75 (MD5)
Made available in DSpace on 2016-01-06T14:43:30Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 1912842 bytes, checksum: dbfd74608eda29f9015d9d5015514d75 (MD5) Previous issue date: 2015-08-14
There are countless social situations in which subjects must take a placement, issue an opinion, present justifications to legitimize the thesis and/or refute contrary opinions. That is, in social life, often, we are called to discursive practice the argumentation, which can be defined as a social activity of fundamentally dialogical nature. The argument writing plays crucial role in the integration of young people in social communicative practices developed by today's society. Given this role, the research that led to this thesis aimed to analyze argumentative essays produced by graduating high school students, checking what are the structural and discursive character of resources that draw students to write the dissertative-argumentative text. The hypothesis that guided the research was that students can build argumentative texts, mobilizing various structural and discursive resources. However, despite this, the produced argumentative essays must overcome certain limitations of the argumentative employed standard, which lacks more sophisticated strategies. Based on this hypothesis, we have set as a general objective to analyze the construction of argumentation in written texts produced by thirty (30) graduating high school students, checking what are the structural and discursive character of resources that draw them to sustain their point of view. As a theoretical basis, we make use of the work of Anscombre and Ducrot (1994), Aristotle (sd), Bakhtin (1988, 1997, 2002), Bunzen (2006), Ducrot (1990, 2009), Marcuschi (1983, 1986, 2002 , 2008), Platin (2005), Perelman and Olbrechts-Tyteca (2005), Reinaldo (2001), Toulmin (2006), Van Dijk (1989, 2010), among other authors. As can be seen by the authors mentioned above, we do not restricted ourselves to a theoretical approach, given the complex nature of our object of analysis. The focus of the analysis was predominantly qualitative. The results we have come to confirms our hypothesis. With regard to the fact that most of the volunteer students can build argumentative texts, mobilizing various structural and discursive resources, it is clear that this is due to: the titles, mostly, of a kind suggestive and mostly argumentative; the appropriate division of paragraphs; the presence of different text blocks (introduction, development and conclusion), even in texts that were not properly divided into paragraphs; the presence of a thesis to be defended. In what concerns the limitations to overcome, we find these are linked to the organization of arguments, since most volunteers build arguments with minimum standard (data, justificative and conclusion), taking us call to mind the low occurrence use of backup and refutation in the arguments. This latter finding is consistent with the findings in relation to the critical positioning, which is discursively constructed by means of the verb-axiological position that is assumed in front of heteroglossia. The subject takes a point of view, as it performs two major dialogical evaluative movements: assimilation and taking distance of other people's speech. The analyzed corpus, In dominated the assimilation of other people's speeches, the main strategy incorporating voices without the delimitation of unrelated sayings. The authors that incorporate others voice turn it into his own voice, because of its critical positioning. We found that the greater the distance, the more explicit becomes the responsibility of the enunciated, and the greater assimilation the less clear becomes the boundaries between speech of the enunciator subject and another's speech
Il existe d'innombrables situations sociales dans lesquelles les sujets doivent prendre un placement, émettre un avis, présenter des justifications qui légitiment la thèse défendue et/ou réfuter les opinions contraires. C’est-à-dire, dans la vie sociale, souvent, nous sommes appelés à la pratique discursive de l'argumentation, qui peut être définie comme une activité sociale fondamentalement de nature dialogique. L'argumentation écrite joue un rôle crucial dans l'intégration des jeunes dans les pratiques sócio-comunicatives développées par la société d'aujourd'hui. Étant donné que, la recherche qui a conduit à cette thèse vise à analyser dissertations argumentatifs produits par les étudiants de la fin du lycée, en vérifiant ceux qui sont les recours de caractère structurel et discursif qui attirent les étudiants à écrire le texte dissertatif-argumentatif. L'hypothèse qui a guidé la recherche était que les étudiants peuvent construire des textes argumentatifs, mobiliser diverses ressources structurelles et discursives. Cependant, malgré cela, les dissertations argumentatives produits doivent surmonter certaines limitations provenants de la norme argumentative employée, qui manque de stratégies plus sophistiquées. Basé sur cette hypothèse, nous avons fixé comme objectif général d'analyser la construction de l'argumentation dans les textes produits par les trente (30) étudiants de la fin du lycée, en vérifiant ceux qui sont les recours de caractère structurel et discursif qui attirent les étudiants à soutenir la défense d’un point de vue. Comme une base théorique, nous nous soutenons sur l'oeuvre de Anscombre et Ducrot (1994), Aristote (s.d.), Bakhtine (1988, 1997, 2002), Bunzen (2006), Ducrot (1990, 2009), Marcuschi (1983, 1986, 2002 , 2008), Platin (2005), Perelman et Olbrechts-Tyteca (2005), Reinaldo (2001), Toulmin (2006), Van Dijk (1989, 2010), entre autres auteurs. Comme on peut le voir par les auteurs mentionnés cidessus, nous ne nous limitons pas à une approche théorique, étant donné la nature complexe de l'objet de notre analyse. L'objectif de l'analyse était essentiellement qualitative. Les résultats que nous avons obtenus, ils ont confirmé notre hypothèse. En ce qui concerne le fait que la plupart des étudiants volontaires réussissent à élaborer des textes argumentatifs, en mobilisant diverses ressources structurelles et discursives, il est clair que cela est dû : aux titres, pour la plupart, d'un genre suggestif et correctement argumentatif; la division appropriée des paragraphes; la présence de différents blocs de texte (introduction, développement et conclusion), même dans les textes qui n’ont pas été correctement répartis en paragraphes; la présence d'une thèse à défendre. En ce qui concerne les limitations à surmonter, nous trouvons qui ceux-ci sont liés à l'organisation d'arguments, puisque la plupart des volontaires construisent des arguments avec la norme minimale (étant donné la justification et la conclusion), en faisant attention à l'esprit de la faible occurrence de soutien et de réfutation aux arguments. Cette dernière constatation est conforme aux conclusions relatives à la position critique, qui est discursivement construit à partir de la position du verbeaxiologique qui est supposé devant hétéroglossie. Le sujet prend un point de vue, car il effectue deux grands mouvements d'évaluation dialogique: assimilation et d’éloignement de la parole des autres. Dans le corpus analysé, il a dominé l'assimilation des discours d'autres personnes, en ayant comme principale stratégie d’intégration des voix sans la délimitation des énonciations non apparentées. Les auteurs, qui en prenant ce discours de l’autre, les transforment en sa propre voix. Nous avons constaté que, plus il y a l’éloignement, plus explicité devient la responsabilité de ce qui parle, et plus grande l’assimilation devient, moins clair sont les frontières entre le discours du sujet qui porte la parole et le discours d’autre.
São inúmeras as situações sociais em que os sujeitos precisam assumir um posicionamento, emitir uma opinião, apresentar justificativas que legitimem a tese defendida e/ou refutar opiniões contrárias. Ou seja, na vida social, frequentemente, somos convocados à prática discursiva da argumentação, que pode ser definida como uma atividade social de natureza fundamentalmente dialógica. A argumentação escrita exerce papel crucial para a inserção dos jovens nas práticas sociocomunicativas desenvolvidas pela sociedade atual. Considerando esse papel, a pesquisa que gerou esta tese se propôs a analisar dissertações argumentativas produzidas por estudantes concluintes do Ensino Médio, averiguando quais são os recursos de caráter estrutural e discursivo de que se valem os alunos para redigir o texto dissertativo-argumentativo. A hipótese que norteou a pesquisa foi a de que os estudantes conseguem construir textos argumentativos, mobilizando variados recursos estruturais e discursivos. No entanto, apesar disso, as dissertações argumentativas produzidas precisam superar certas limitações decorrentes do padrão argumentativo empregado, que carece de estratégias mais sofisticadas. Com base nessa hipótese, estabelecemos, como objetivo geral, analisar a construção da argumentação em textos escritos produzidos por 30 (trinta) estudantes concluintes do Ensino Médio, averiguando quais são os recursos de caráter estrutural e discursivo de que se valem os alunos para sustentar a defesa de um ponto de vista. Como base teórica, valemo-nos dos trabalhos de Anscombre e Ducrot (1994), Aristóteles (s.d.), Bakhtin (1988, 1997, 2002), Bunzen (2006), Ducrot (1990, 2009), Marcuschi (1983, 1986, 2002, 2008), Platin (2005), Perelman e Olbrechts-Tyteca (2005), Reinaldo (2001), Toulmin (2006), Van Dijk (1989, 2010), entre outros autores. Como é possível constatar pelos autores anteriormente citados, não nos restringimos a um enfoque teórico, dada a natureza complexa do nosso objeto de análise. O enfoque da análise foi predominantemente qualitativo. Os resultados a que chegamos confirmam nossa hipótese. No que diz respeito ao fato de a maioria dos estudantes voluntários conseguirem construir textos argumentativos, mobilizando variados recursos estruturais e discursivos, percebe-se que isso se deve aos seguintes aspectos: os títulos, em sua maioria, dos tipos sugestivos e propriamente argumentativos; a adequada divisão dos parágrafos; a presença de distintos blocos textuais (introdução, desenvolvimento e conclusão), mesmo nos textos que não foram corretamente divididos em parágrafos; a presença de uma tese a ser defendida. Já no que diz respeito às limitações a serem superadas, verificamos que estas estão ligadas à organização dos argumentos, uma vez que a maioria dos voluntários constrói os argumentos com padrão mínimo (dado, justificativa e conclusão), tendo-nos chamado à atenção a baixa ocorrência de uso de apoio e de refutação nos argumentos. Essa última constatação se coaduna com os achados em relação ao posicionamento crítico, que é discursivamente construído por meio da posição verbo-axiológica que se assume frente à heteroglossia. O sujeito assume um ponto de vista, à medida que realiza dois grandes movimentos dialógico-valorativos: a assimilação e o afastamento do discurso alheio. No corpus analisado, predominou a assimilação de discursos de outrem, tendo como principal estratégia a incorporação de vozes, sem a delimitação de dizeres alheios. Os autores que incorporam o dizer alheio o transformam em sua própria voz, em função de seu posicionamento crítico. Constatamos que, quanto maior o afastamento, mais explícita se torna a responsabilidade de quem enunciou, e quanto maior a assimilação, menos nítidas se tornam as fronteiras entre o discurso do sujeito enunciador e o discurso de outrem.

We use a Fully Convolutional Neural Network (FCNN) to classify pixels in historical document images, enabling the extraction of high-quality, pixel-precise and semantically consistent layers of masked content. We also analyze a dataset of hand-labeled historical form images of unprecedented detail and complexity. The semantic categories we consider in this new dataset include handwriting, machine-printed text, dotted and solid lines, and stamps. Segmentation of document images into distinct layers allows handwriting, machine print, and other content to be processed and recognized discriminatively, and therefore more intelligently than might be possible with content-unaware methods. We show that an efficient FCNN with relatively few parameters can accurately segment documents having similar textural content when trained on a single representative pixel-labeled document image, even when layouts differ significantly. In contrast to the overwhelming majority of existing semantic segmentation approaches, we allow multiple labels to be predicted per pixel location, which allows for direct prediction and reconstruction of overlapped content. We perform an analysis of prevalent pixel-wise performance measures, and show that several popular performance measures can be manipulated adversarially, yielding arbitrarily high measures based on the type of bias used to generate the ground-truth. We propose a solution to the gaming problem by comparing absolute performance to an estimated human level of performance. We also present results on a recent international competition requiring the automatic annotation of billions of pixels, in which our method took first place.

Cette thèse porte sur l’influence de l’anglais comme langue mondiale sur la presse numérique grecque. C’est une étude qualitative d’un corpus de 126 textes de la presse grecque et d’un corpus complémentaire de 15 textes de la presse anglophone publiés en ligne entre 2011-2015. Les données sont tirées de trois genres, les articles d’actualité, les articles d’opinion et les interviews, et cinq thématiques : art et mode, économie et politique, monde et environnement, science et technologie et santé et sport. Le cadre théorique s’appuie sur l’emprunt linguistique et la rhétorique d’un texte numérique. Quant à l’influence de l’anglais, nous notons des traces au niveau du lexique, de la syntaxe, de la sémantique, du discours et de la titraille. Pour la mise en scène d’un texte numérique, nous notons divers modèles de lecture, ainsi que les différences entre la presse grecque et anglophone. Finalement, ces résultats soulignent une combinaison de l’implantation de l’anglais dans la presse grecque du haut vers le bas et du bas vers le haut
This thesis focuses on the influence of English as a global language on the Greek digital press. It is a qualitative study of a corpus of 126 texts from the Greek online press and a complementary corpus of 15 texts from the anglophone press published on the Internet between 2011-2015. The data are drawn from three journalistic genres: news articles, opinion articles and interviews, and 5 column-themes: art and fashion, economy and politics, world and environment, science and technology and health and sports. The theoretical framework is based on the linguistic borrowing and the rhetorical theory of a digital text. As for the influence of English, traces on the lexicon, syntax, semantics, discourse and headlines are mainly pointed out. Concerning the layout of a digital text, we found various reading models, as well as some differences between the Greek and anglophone press. Finally, these results underline a combination of the implantation of English borrowing on the Greek press from a top down and a bottom up point of view

Un nombre incalculable de documents est imprimé, numérisé, faxé, photographié chaque jour. Ces documents sont hybrides : ils existent sous forme papier et numérique. De plus les documents numériques peuvent être consultés et modifiés simultanément dans de nombreux endroits. Avec la disponibilité des logiciels d’édition d’image, il est devenu très facile de modifier ou de falsifier un document. Cela crée un besoin croissant pour un système d’authentification capable de traiter ces documents hybrides. Les solutions actuelles reposent sur des processus d’authentification séparés pour les documents papiers et numériques. D’autres solutions reposent sur une vérification visuelle et offrent seulement une sécurité partielle. Dans d’autres cas elles nécessitent que les documents sensibles soient stockés à l’extérieur des locaux de l’entreprise et un accès au réseau au moment de la vérification. Afin de surmonter tous ces problèmes, nous proposons de créer un algorithme de hachage sémantique pour les images de documents. Cet algorithme de hachage devrait fournir une signature compacte pour toutes les informations visuellement significatives contenues dans le document. Ce condensé permettra la création de systèmes de sécurité hybrides pour sécuriser tout le document. Ceci peut être réalisé grâce à des algorithmes d’analyse du document. Cependant ceux-ci ont besoin d’être porté à un niveau de performance sans précédent, en particulier leur fiabilité qui dépend de leur stabilité. Après avoir défini le contexte de l’étude et ce qu’est un algorithme stable, nous nous sommes attachés à produire des algorithmes stables pour la description de la mise en page, la segmentation d’un document, la reconnaissance de caractères et la description des zones graphiques
An innumerable number of documents is being printed, scanned, faxed, photographed every day. These documents are hybrid : they exist as both hard copies and digital copies. Moreover their digital copies can be viewed and modified simultaneously in many places. With the availability of image modification software, it has become very easy to modify or forge a document. This creates a rising need for an authentication scheme capable of handling these hybrid documents. Current solutions rely on separate authentication schemes for paper and digital documents. Other solutions rely on manual visual verification and offer only partial security or require that sensitive documents be stored outside the company’s premises and a network access at the verification time. In order to overcome all these issues we propose to create a semantic hashing algorithm for document images. This hashing algorithm should provide a compact digest for all the visually significant information contained in the document. This digest will allow current hybrid security systems to secure all the document. This can be achieved thanks to document analysis algorithms. However those need to be brought to an unprecedented level of performance, in particular for their reliability which depends on their stability. After defining the context of this study and what is a stable algorithm, we focused on producing stable algorithms for layout description, document segmentation, character recognition and describing the graphical parts of a document

Cette thèse de doctorat se consacre à l’analyse de périodiques créés au cours du XXe siècle par des cinéastes et des plasticiens à l’œuvre dans le champ des avant-gardes et du cinéma expérimental. Les revues forment des objets plastiques et spéculatifs, complexes et composites de par les relations qui se nouent entre le texte et l’image, les montages qui se créent et le défi que constitue la transposition des images filmiques. En quoi ces revues d’artistes témoignent-elles d’une recherche expérimentale ? Comment les revues d’artistes participent-elles à une histoire critique et plastique des formes cinématographiques ? L’étude tente de comprendre les manières originales dont les cinéastes et les plasticiens se saisissent des revues afin d’élaborer, défendre, documenter, objectiver et analyser certains paradigmes cinématographiques. À quels titres les revues deviennent elles-mêmes des propositions expérimentales, des laboratoires de recherche sur les liens entre l’image et le texte ? Nous observerons comment, grâce à leurs propositions techniques, graphiques et visuelles propres, les revues exposent certains enjeux matériels, poétiques, plastiques et théoriques propres à l’image cinématographique, comment elles questionnent le regard. Les revues offrent des plateformes de diffusion et de dissémination esthétiques, servent à ouvrir des réseaux de circulation pour les idées, singulières ou collectives, des rédacteurs en chef. Comment accompagnent-elles leurs efforts dans la construction d’un milieu cinématographique alternatif ? Les revues Dada I de Tristan Tzara et Hans Arp (1916), Dada Sinn der Welt de John Heartfield et George Grosz (1921), Le Promenoir de Jean Epstein, Pierre Deval et Jean Lacroix (1921-1922), G. für elementare Geschaltung de Hans Richter (1923-1926), Close Up du groupe Pool composé de Kenneth Macpherson, Bryher et H.D. (1927-1933), Film Culture de Jonas Mekas (1955-1996) et Cantrill’s Filmnotes d’Arthur et Corinne Cantrill (1971-2000) forment le corpus de cette thèse qui vise à contribuer à une histoire plastique des publications expérimentales
This PhD thesis focuses on analyzing periodicals created during the XXth Century by both visual artists and filmmakers operating in the realm of avantgardes and experimental cinema. The journals become plastic, conceptual, complex, and composite objects because of the interplay between text and image as well as the reproduction of images and realization of photomontages. How these artists’ journals show signs of an experimental approach ? How do artists’ journals contribute to the critical and plastic history of film ? The dissertation aims to understand the unique ways the visual artists and filmmakers make use of the journals to create, defend, document, visualize and analyze some cinematic paradigms. To what extent the journals become in turn experimental works about the relationships between text and image ? We will study how magazines exhibit various plastic, aesthetical, theoretical, and poetical dimensions at stake in the cinematic image, relying on specific technical, graphic and visual undertakings, and how they call into question the perception. Journals become instrumentalized in ensuring the movement of the editors’ ideas, either collective or indivuals. How do journals support the editors’ efforts in building an alternative cinema domain ? Dada I edited by Tristan Tzara and Hans Arp (1916), Dada Sinn der Welt by John Heartfield and George Grosz (1921), Le Promenoir by Jean Epstein, Pierre Deval and Jean Lacroix (1921-1922), G. für elementare Geschaltung by Hans Richter (1923-1926), Close Up by Kenneth Macpherson, Bryher and H.D. (1927-1933), Film Culture by Jonas Mekas (1955-1996) and Cantrill’s Filmnotes by Arthur et Corinne Cantrill (1971-2000) form the corpus of this PhD thesis, which aims to contribute to a plastic history of experimental publications

碩士
國立臺北科技大學
工業工程與管理研究所
92
Association Rule Mining aims to discover the customers’ buying patterns in order to assist in product layout decisions and sales promotion decisions. Traditional association rules consider only the conjunctive associations among items. In practice, other associations among items, such as disjunctive associations, may be very useful for decision making. In this thesis, we consider the use of the generalized association rules for page layout. We show that the generalized association rules can be very useful for designing the commercial websites and performing one-to-one marketing. An example system that uses the generalized association rules for courses recommendations for students is developed and reported.

With contemporary research focusing its attention primarily on benchmark-driven performance evaluation, studying fundamental memory characteristics has gone by the way-side. This thesis presents a systematic study of the expected performance characteristics for contemporary multi-core CPUs. These characteristics are the primary influence on benchmarking variability and need to be quantified if more accurate benchmark results are desired. With the aid of a new, highly customizable, micro-benchmark suite, these CPU-specific attributes are evaluated and contrasted. The benchmark tool provides the framework for accurately measuring instruction throughput and integrates hardware performance counters to gain insight into machine-level caching performance. Additionally, the Linux operating system's impact on cache utilization is evaluated. With careful virtual memory management, cache-misses may be reduced, significantly contributing to benchmark result stability. Finally, a popular cache performance model, stack distance profile, is evaluated with respect to contemporary CPU architectures. While particularly popular in multi-core contention-aware scheduling projects, modern incarnations of the model fail to account for trends in CPU cache hardware, leading to measurable degrees of inaccuracy.

碩士
大同大學
工業設計研究所
90
Advances in technology have benefited mankind with many conveniences in their day-to-day activities, as evidenced by the rapid rise in the number of people who use the World Wide Web to search for information that they need. For these people, the need for a quick and easy way to retrieve the latest information has become a key requirement. With more and more people turning to the small screen for instant messaging, this need has now become the driving force for the growing popularity of portable computer technology. Although the PDA has enjoyed wide popularity as a convenient and portable information retrieval instrument, it has its limitations in that its small screen size cannot accommodate the same amount of data as a desktop computer, causing much inconvenience to the user. Although some usability guidelines have been developed for large screen, few guidelines are available for ensuring the usability of Web pages on PDA platform. The issue of providing the user with convenient and clear presentation of data during the search for information is one that must be carefully studied. Our research assumes that, with the restricted size of the small PDA screen, the screen display of data collected from various shopping websites is affected by the layout of the content on the web page and the navigational functionality of the web page. This in turn impacts the search time of the PDA user and the effectiveness of this information to the user. First, we used a PDA to retrieve and display the web pages of a collection of shopping websites and compared the screen displays to those of non-PDA screens. And then design the experiment web page and test it. In Phase I, the results showed that varying the levels of the 3 design factors （content layout, site and within-page navigation） did not have any distinct impact on the web pages without frames. In Phase II, the results showed that anchoring the navigational tool within a frame impacts the effectiveness of the user during online shopping. In summary, the results also showed that the ideal web layout is “pictures followed by text” with site navigational tool located on the top of the frame and within-page navigational tool located on the left hand side of the frame. Our research results can serve as a basis for web page layout design on PDA platform. Keywords: Personal Digital Assistant (PDA), Product Exhibition, Content Layout, Navigation

碩士
大同大學
工業設計學系(所)
102
The purpose of this study is to investigate the emotion of page layout in mobile device. To do this, this study planned a pilot study and three experiments. 　　The pilot study collected the top 100 APPs in the free ranking and paid ranking of the Google Play and App Store.The page layouts collected were analyzed. The results of pilot study were used for Experiment 1 and 2. In order to understand the difference between the different sizes of mobile devices, Experiment 1 used different device sizes showing the same page layout to compare with each other. Experiment 2 investigated the emotions of page layouts. Experiment 3 used interactive page layouts to investigate the relationship between the emotions of page layouts and the completed time of tasks. The major findings are summarized below: 1.The existing layouts can be classified into five styles, they were “one column grid,” “modular grid,” “variation grid,” “hierarchical grid,” and “tab grid”. 2.The influences of mibile device size and image types on emotions were found insgnificant. 3.The emotion space was constituted by “trust” factor and “innovation” factor. 4.The page layouts were distinguished into 6 groups. They were “outstanding group,” “nervous group,” “order group,” “boring group,” “bold group,” and “impressive group”. 5.The page layouts having fewer words, fewer images, and fewer horizontal grid quantities were found to have “stable” emotion. 6.The page layouts having fewer proportions of the white background and more horizontal grid quantities were found to have innovation emotion. 7.The emotions judged by female observers tended to “like,”and “friendly” emotions, male “dislike,”and “unfriendly” emotions. 8.The emotions judgment made by the observers who changed the ringtone of the mobile devices tended to “fashion,” “elegant,” “friendy,”and “satisfying” emotions. 9.The emotions judgment made by the observers who changed the buttom bar of the mobile devices tended to “outdated,” “vulgar,” “veiled,” “unfriendly,”and “expected” emotions. 10.The emotions judgment made by the observers who changed the lockscreen of the mobile devices tended to “like,” “interesting,”“fashion,” “elegant,” “friendy,” “trusty,” “satisfying,” “relaxed,”and “surprising” emotions. 11.The emotions judgment made by the observers who have less experience tended to “like,” “interesting,”“fashion,” “elegant,” “bold,” “friendy,” “trusty,” “satisfying,” “relaxed,”and “surprising” emotions. 12.The completed time on interative page were found to be shorter on the page without images than those with images. 　　The result obtained in this study successfully developed the design suggestion, intending to assist designer to design the page layout in mobile device.

碩士
國立臺灣大學
工程科學及海洋工程學研究所
104
In recent year, the mobile device has become the most common tool to access the Internet. In response to the flourish of the mobile device, Ethan Marcotte proposed a design guide which is called Responsive Web Design (RWD) [36] to make the mobile service user-friendly. According to Google statistics, 74 percent of mobile device users prefer user-friendly websites due to its readability on mobile devices. Google Search Guide [31] also reported that a user-friendly website can improve its search ranking and attract users. However, 85 percent of the websites are still user-unfriendly, such that these websites obtain the lower and lower search ranking. The major dilemma is that website rebuilding is usually constructed manually, which is time-consuming and inefficient. If a system can automatically select an appropriate website template, it will significantly improve the efficiency of website rebuilding. The critical problem of this system is how to select an appropriate template which has more features that its customer needs. The study [2] indicates that drastic changes of visual appearance of Web pages have a negative influence on readers. In order not to affect user-experience, this thesis proposes a system that will efficiently sort out the templates which are similar with respect to their layout features and visual appearance. Our experimental results indicate the effectiveness of our approach and show that our approach can find the similar templates precisely. In this thesis, we focus on discussing issues related to the similarity of the Web pages. Since a website consists of multiple Web pages, the proposed method can be extended to measure the similarities between “Websites”.

碩士
國立臺灣科技大學
設計研究所
101
The tools and techniques used in this research were the combination of HyperText Markup Language (HTML) and Cascading Style Sheets (CSS). A method was developed to enable web page designers to control web page layout and to reduce preview differences of web pages to the minimum. Through the techniques of CSS, the result could help web page designers and website developers with more control over web page layout and design in less time and easier to edit and maintain web pages. This research discuss possible solutions through literature reviews and experiments. A standard principle was proposed for cross-browser compatible design in five steps: structuring website layout, declaring DOCTYPE, resetting defaults, applying the box model principle, regulating and verifying for W3C. The prior experiment revealed subjects’ current problems and the formal experiment justified that the proposed standard principle could solve cross-browser problems effectively and could reduce browsing differences of web pages. Designers can use this integral method to reduce preview differences among browsers. It can be a bridge among visual design, page layout and program design. It can increase designers’ working efficiency in designing websites. It also complies with the trend of SEO keyword searching rules to increase website visibility. Such a method of reducing differences is important in future web accessibility and can be extended to different versions of browsers and even on mobile devices. The ideal is duly reflected in this research.

碩士
國立中正大學
資訊管理所
95
With the rapid development of the World Wide Web (WWW), spectacular ascent in size and popularity of information are published on the Internet. This makes information-overload problem from bad to worse. It becomes more complicate to deal with the enormous information space on the WWW. Hence, web classification is the essential and primary task for manage information on the Internet. We develop a web page layout weighting scheme for HTML pre-process to deal with degree of importance of information from user’s visual perception point of view. By mapping from position, size, and depth of a web page to a degree of freedom, the layout weighting scheme generates a continuous right skew distribution to estimate the importance of a block on a web page of a website, so that we can weight importance of information components in a block for making classification more accuracy. Along with this layout weighting scheme, multi-class support vector machine classifiers are applied to implement classification architecture for websites classification. SVM can only deal with binary classification problem. This is oversimplified, since website content may be involved in more than one category. It can be categorized into multiple classes. Therefore, a multi-class classifier is needed for classification. We consider that selection of multi-class SVM and kernel functions should be different due to the different application domains. In this research work, we proposed a layout weighting scheme and with different kernel functions used by multi-class support vector machine classifiers in experiments in order to find the best performance approach. Experiments show that our web page layout weighting scheme improve both classification accuracy and percentage of support vectors. In addition, experimental results suggest that, among multi-class SVM classifiers, One-Against-One classifier is the best for websites classification. Classifiers with RBF kernel demonstrate higher accuracy and F1 score, while polynomial kernel produces lower percentage of support vectors.

碩士
元智大學
資訊傳播學系
99
The rapid development of Internet in recent years has offered us with a diversification of reading choices. The web pages are able to transmit large numbers of real-time information to users through the Internet platform. However, as the current composition and layout of web page design are lacked of specific and precise research guidelines for references, they often cause the designers to use the intuitive customary habit in designing and this prone to missing the opportunity to seize users’ attention. Therefore, how to lay out the web pages well and convey the integrated messages to users clearly is an important issue in web page design. This study retrieved 78 outstanding samples from Design Charts website to serve as the analytical samples. The impact of emotional awareness and preferences on different web page compositions on users were measured, and proposals were made on “main visual position,” “graphic and text proportion,” “blank capacity,” “overall composition layout,” and other layout elements were cross-examined through a content analysis and a questionnaire survey to work as the design principles of web page design. The study discovered that: (1) The composition with “main visual position at the screen center” was able to enhance the “special feeling” of users on screen image effectively. (2) The composition proportion of “diagram bigger than text” was able to improve the “easy feeling” of users on reading the screen. (3) A blank space of “40%~60%” was able to upgrade the “simplicity feeling” of users while watching the screen, but there was still room for further exploration on the effectiveness. (4) The overall composition layout of “main visual position at the center, diagram bigger than text, and blank space larger than 40%” were able to improve the “popular sense” and “love feeling” of users on the overall layout of web pages.

碩士
元智大學
資訊傳播學系
98
Style analysis is essential in researches of product design and arts. However it is rare in commercial and planar design. Newspaper is selected as the research topic in this thesis. Statistical methods such as quantitative analysis is used to analyze three major newspapers in Taiwan. The results are concluded as follows (1) Style analysis and comparison for each period of newspaper history: quantitative analysis is used to analyze the styles of the three major newspapers. The results are compared to references discussing historical background and design trends. The analysis is composed of (a) style difference before and after censorship on the press, (b) style difference before and after the release of "Apple Daily", (c) style difference between the three major newspapers and "Apple Daily" before its release, (d) style difference between the three major newspapers and "Apple Daily" after its release and (e) comparison of the three major newspapers and "Apple Daily". (2) Style generation of newspaper layouts: The analyzed data are used to generate newspaper layouts to verify the results.

碩士
國立臺灣藝術大學
應用媒體藝術研究所
93
This thesis explores the construction of professional competency indices on page layout for trainees at vocational training. The purpose of this study is to (1) identify the professional competency of vocational trainers required by the vocational training of page layout; (2) recognize the necessary professional competency of vocational trainees identified by the experts in the printing industry; (3) distinguish the difference of necessary competency between vocational training and printing industry; (4) construct the professional competency indices on page layout for trainees at vocational training. The methodologies of literature review and questionnaire survey are utilized in this thesis. The research emphasizes on the data collected from both vocational page layout trainers and the experienced experts in the printing industry. There were 360 random selected samples were surveyed, and 182 valid questionnaires were collected (an effective return rate of 50.56%). The results of this study were found by analyzing the collected data through Frequency Distribution, Percentage, Mean Comparison, t-test, One-way ANOVA, and Scheffé method. Based on the purpose of this study, several findings are concluded as follows: (1) With the analysis from questionnaire survey, a vocational trainee should be capable of 44 professional competency under three categories; including 23 very important professional competency and 26 important professional competency. (2) There were no significant differences, between vocational trainers and industry experts, in professional competency indices on page layout for the vocational trainees. (3) There were no significant differences in professional competency between male and female, which include various educational backgrounds, seniorities, geographic areas, and teaching subjects. The differences were computed by Scheffe Method and summarized in conclusions. Finally, this research proposes practical conclusions and recommendations as references for course design and future research in the area of page layout to the managing division of vocational training（Bureau of Employment and Vocational Training）, public and private vocational training institutions, as well as training organizations authorized by Bureau of Employment and Vocational Training.

碩士
國立雲林科技大學
資訊管理系碩士班
94
Information overload is a generally a problem occurring in the whole Internet world, but also a serious problem on a single web page. Too much information will result in user having to search too long for data. A good design, which facilitates a person’s web navigation behavior, would generally lead to better user performance. Web menus which can present information structure effectively have become a tool to resolving the information overload problem of a single web page. In these days, it is common to see that there are two menus on one web page. To understand which menu type and menu arrangement is the best strategy to support the navigator’s work, this study investigated the commonly used menu types and arrangement strategies. There are two types of menus: the event trapping menu and the simultaneous hierarchical menu. These menus are generally located on the top and left position of the web page. Combining the menu type and position variable, there are four menu arrangement strategies to be the testing interface. 24 participants were invited to perform searches using these four strategies. Performing time and subjective satisfaction are the independent variables measured during interface testing. The results show that there is no significant difference in these four menu arrangements in user performance. In the user subjective satisfaction, however, there is a significant difference on several questions. To sum up, the event trapping menu using the top and left position is the best menu arrangement for supporting the user searches. On the other hand, the simultaneous hierarchical menu using the top and left side is not an ideal menu arrangement for users doing web searches.

碩士
國立臺北科技大學
工業工程與管理研究所
98
A complex shopping website often confuses users. A poor web design refrains the user from obtaining the desired information. Current academic research lacks a criterion for the home page layout design. This study offers suggestions to help web designers on the design of web page block re-layout. According to proximity compatibility principle, a mental compatibility and the integration of attention, a better design by re-layout or re-display web page blocks with a cluster closed each other can reduce waste of scanning and attention, and ultimately, it improves the readability of the webs. Yahoo Shopping Website is the most common shopping website in Taiwan, so it was used as the case in this study for redesign. The card sorting method was utilized to identify subjects'' metal models on the proximities of home page blocks on the shopping web portals. Two independent variables, block layout and color, were design in the experiment. Subjects'' eye movement was collected and analyzed through a gaze tracking system. Results showed that the best design was the re-layout of the page blocks. The findings of this study suggest that using the Proximity Compatibility Principle for designing web pages should be able to reduce the waste time of obtaining information and enhance browsing efficiency.

碩士
國立雲林科技大學
視覺傳達設計系
103
Electronic commerce is an entry interface of hyperlinks for presenting a website’s information to visitors, which is mostly concerned with human perception in terms of users’ comprehension and mental demands. The visual effect of web banner and content advertisements are especially important in electronic commerce interface design and assume a marketing strategy. With the prosperous growth of electronic commerce, cascading style sheets framework supply designers the structure of website and modeling template, website towards cross-platform and be used extensively in other web devices. This paper aims to derive website page styles from design factors of essential web framework and accord with the demands of the market, and pursue to provide a model for web image design which meets customers’ needs. First, semantic differential Technique is used to investigate the influence of visual elements on website banner and content framework and then we present kansei engineering analysis of visual components properties of cognitive. Finally, multivariate analysis of variance method is used to select an optimal design strategy. The research presents a visual interface design approach for web developers to reduce cost and proposes an optimal decision making deal with a website image design. The results of website page design model is applicable to other forms of cross-platform interface design contributes the electronic commerce website design.

碩士
銘傳大學
資訊管理學系碩士在職專班
100
It is undeniable that the Internet has become the most important pathway for people to gain information nowadays. Browsing web pages has become everyday life. Internet is getting more and more popular, and the number of web pages is growing rapidly. Under these circumstances, it is unavoidable to have competitor websites for Electronic Business owners; therefore, every Electronic Business owners and web pages designers fix their eyes on how to attract users to browse, stay in and come back to their websites. In order to achieve those goals, the websites need to have abundant and rapidly-updated contents. Meeting users’ sensory preferences in visuals design is also the key. Web pages designers try their best at using different color combinations and frames to enhance sensory comfort level. However, different business model could provide different user experience and different value, but few researcher focuses on distinguishing the impacts that different business models may have. This research tries to explore the impacts of web page layout and it’s color combinations on users’ Perception via Legibility. Furthermore, we would like to answer the question whether the interaction of layouts and color combinations have moderating effects on Legibility in Portal sites, News and Social networks. Our research uses the 3x2x2 factor analysis. Researcher developed an experimental websites and sampling from collected data. Valid samples total is 936. The result shows that an opposite L style layout has more positive comments on Legibility than menu-top style layout. The research also finds out that among the five color combinations, the more difference of color level of color combinations or the minus result of color level of color combination is zero, the higher scores on Legibility, indicating a preference of strong and contrasted color combinations. Moreover, websites layout and color combinations combines together helps moderate Legibility. News websites are effected the most, Portal sites next, and the last Social networks. Therefore we can conclude that websites layout and color combinations do make a difference in Legibility. Website designers have to follow certain rules when choosing color combinations, and to consider the layouts and designing elements for different websites in order to create popular websites. At last, apart from contributing in theory and in practice, this research also points out some research limitations and suggestions for further investigation.

A document image, beside text, may contain pictures, graphs, signatures, logos, barcodes, hand-drawn sketches and/or seals. Further, the text blocks in an image may be in Manhattan or any complex layout. Document Layout Analysis is an important preprocessing step before subjecting any such image to OCR. Here, the image with complex layout and content is segmented into its constituent components. For many present day applications, separating the text from the non-text blocks is sufficient. This enables the conversion of the text elements present in the image to their corresponding editable form. In this work, an effort has been made to separate the text areas from the various kinds of possible non-text elements. The document images may have been obtained from a scanner or camera. If the source is a scanner, there is control on the scanning resolution, and lighting of the paper surface. Moreover, during the scanning process, the paper surface remains parallel to the sensor surface. However, when an image is obtained through a camera, these advantages are no longer available. Here, an algorithm is proposed to separate the text present in an image from the clutter, irrespective of the imaging technology used. This is achieved by using both the structural and textural information of the text present in the gray image. A bank of Gabor filters characterizes the statistical distribution of the text elements in the document. A connected component based technique removes certain types of non-text elements from the image. When a camera is used to acquire document images, generally, along with the structural and textural information of the text, color information is also obtained. It can be assumed that text present in an image has a certain amount of color homogeneity. So, a graph-theoretical color clustering scheme is employed to segment the iso-color components of the image. Each iso-color image is then analyzed separately for its structural and textural properties. The results of such analyses are merged with the information obtained from the gray component of the image. This helps to separate the colored text areas from the non-text elements. The proposed scheme is computationally intensive, because the separation of the text from non-text entities is performed at the pixel level Since any entity is represented by a connected set of pixels, it makes more sense to carry out the separation only at specific points, selected as representatives of their neighborhood. Harris' operator evaluates an edge-measure at each pixel and selects pixels, which are locally rich on this measure. These points are then employed for separating text from non-text elements. Many government documents and forms in India are bi-lingual or tri-lingual in nature. Further, in school text books, it is common to find English words interspersed within sentences in the main Indian language of the book. In such documents, successive words in a line of text may be of different scripts (languages). Hence, for OCR of these documents, the script must be recognized at the level of words, rather than lines or paragraphs. A database of about 20,000 words each from 11 Indian scripts1 is created. This is so far the largest database of Indian words collected and deployed for script recognition purpose. Here again, a bank of 36 Gabor filters is used to extract the feature vector which represents the script of the word. The effectiveness of Gabor features is compared with that of DCT and it is found that Gabor features marginally outperform the DOT. Simple, linear and non-linear classifiers are employed to classify the word in the feature space. It is assumed that a scheme developed to recognize the script of the words would work equally fine for sentences and paragraphs. This assumption has been verified with supporting results. A systematic study has been conducted to evaluate and compare the accuracy of various feature-classifier combinations for word script recognition. We have considered the cases of bi-script and tri-script documents, which are largely available. Average recognition accuracies for bi-script and tri-script cases are 98.4% and 98.2%, respectively. A hierarchical blind script recognizer, involving all eleven scripts has been developed and evaluated, which yields an average accuracy of 94.1%. The major contributions of the thesis are: • A graph theoretic color clustering scheme is used to segment colored text. • A scheme is proposed to separate text from the non-text content of documents with complex layout and content, captured by scanner or camera. • Computational complexity is reduced by performing the separation task on a selected set of locally edge-rich points. • Script identification at word level is carried out using different feature classifier combinations. Gabor features with SVM classifier outperforms any other feature-classifier combinations. A hierarchical blind script recognition algorithm, involving the recognition of 11 Indian scripts, is developed. This structure employs the most efficient feature-classifier combination at each individual nodal point of the tree to maximize the system performance. A sequential forward feature selection algorithm is employed to. select the most discriminating features, in a case by case basis, for script-recognition. The 11 scripts are Bengali, Devanagari, Gujarati, Kannada, Malayalam, Odiya, Puniabi, Roman. Tamil, Telugu and Urdu.

碩士
國立政治大學
傳播學院碩士在職專班
96
This study aims to investigate the relationship between the layout and browsing routes for merchandise display pages on Taiwan’s three major e-commerce websites, as ascertained by results from eye tracking experiments and expert interviews. The study offers some overall conclusions and suggestions which may serve as a reference for future practices in designing and implementing e-commence websites. The experiment was divided into two stages, and included 23 participants whose viewing behavior was recorded; the effective sample pool was 20. The objective of the first test was to examine the merchandise viewing efficacy rendered by these three e-commerce websites. Through this experiment, the website with the merchandise page setup and browsing routes which enabled users to locate their desired merchandise in the shortest time was determined, as well as the samples’ visual fixation distributions and durations. In the second experiment, merchandise page layouts were tested. Different arrangement schemes of merchandise displayed by these same three websites were manipulated to test the effects they have on user eye fixation count and fixation time, in an attempt to understand what type of merchandise display arrangement scheme draws the most attention from viewers and encourages them to spend more viewing time. Expert interviews provided insights and suggestions from skilled web page designers and website builders regarding their perspectives on merchandise display page arrangement and user browsing routes, and in particular, the concepts employed when building their own e-commerce websites. Finally, interview data is introduced and analyzed using content analysis methodology, which may serve as a reference for e-commerce website design practices.

碩士
國立成功大學
資訊工程學系碩博士班
98
Today the Web has become the largest information source for people. Web page contents have been more informative. A web page usually contains various contents such as navigation, interaction and information. When a user is browsing a new web page, he usually needs to spend a lot of time on adopt to the layout of the web page. Bernard, 2002 shows that users always expect that certain functional part of a web page (e.g. navigational links, advertisement bar) appears at certain position of that page. Build a customization template of a category of web pages can improve user browsing experience. In order to build the layout of a web page, we proposed a model to identify the types of blocks. The types are based on the need behind the user, i.e. navigational, informational and transactional plus our observation on web pages, i.e. advertising and social. We use various features to identify the type of a block. The major feature is the CSS selectors extracted from a block. The CSS Selectors control how to display a HTML element and usually have semantic relation with the block. In our observation, there are 52% such CSS Selectors in web pages. We proposed the CSS Selector model to extract the keywords from the CSS Selectors. The model uses three methods, i.e. dictionary look up, probability of n-gram sequence, and branching entropy, to segment a sequence of CSS selectors into keywords. We also use features like hypertexts, sizes of HTML elements and context to help classifying the blocks. In block classification, we proposed a 2-stage method to classify the blocks. The method first predicts a type of a block by CSS Selector Model and combines the result from CSS Selector Model into SVM classifier. The experiments show that our CSS Selector Model can extract relative keywords about the block’s type. It is the most important feature in our classification. The results of block classification can be used to build the layout of web page.

碩士
大同大學
工業設計學系(所)
94
Abstract In face of the competitive work environment and the more and more advanced Internet technology, many designers have started acquire knowledge about using computer graphics software through online forums in recent years. However, as most of the online forums are not well-organized, and the difference of webpage presentation on different computers is not considered, inconvenience in using and learning has been resulted. Thus, many users are unable to be benefited by the advantages of online forums. The objective of this study was to investigate the difference among factors of web layouts for existing online forums. Based on the user groups of different cognitive styles, this study attempted to find out an optimal web layout and set up the design principles for forum webpage as a reference for designers and engineers to construct this type of websites in the future. First of all, according to the literature review, 10 factors of web layouts were set up. From the current online forums, this study collected 10 forums dedicated to the discussion of computer graphics software. Finally, based on the graphics software forums, an experiment was conducted on the web layouts. Through the experiment on webpage components, the analysis results showed a significant difference in white space, page division, contrast between foreground and background, and color among the factors of forum layout components. In other words, research subjects’ satisfactions with white space, page division, contrast between foreground and background, and colors were bipolar. In the aspect of white space, the difference in the proportion of white space columns was the largest. In the aspect of page division, the difference in the location of basic data and utilization of color blocks is the largest. In the aspect of foreground and background contrast, the difference in the contrast of font and background is the largest. Finally, in the aspect of colors, the difference in the number of colors applied for text columns is the largest. In addition, based on the experimental results, this study set “cognitive style”, “location of basic data”, and “utilization of color blocks” as the factors for the experiment on layout arrangement. In the experiment, Group Embedded Figure Test, GEFT, was applied to group the subjects by cognitive styles. Through computer experiment, this study further investigated the impact of different web layouts on users with different cognitive styles. The analysis indicated that users with field independence preferred the placement of basic data on the left, joined published articles with the use of different color blocks, while users of field dependence preferred the placement of basic data on the left, un-joined published articles, and the use of same color blocks.