A tese tem como objetivo o estudo do texto da cidade de Moscou, sua análise comparativa e diálogo com o texto de São Petersburgo, bem como a importância de ambos os textos no processo de formação de identidade cultural russa. A metodologia da pesquisa baseia-se no conceito de texto urbano elaborado por destacados estudiosos da escola semiótica de Tártu-Moscou, tais como: Iúri Lótman, Vladimir Toporov, entre outros. O trabalho apresentado é constituído de três partes: No primeiro capítulo é apresentada a análise teórica dos conceitos de texto e cidade e sua aplicação à Moscou; O segundo capítulo representa uma análise histórico-cultural do processo de formação do texto de Moscou desde a fundação da cidade até o século XXI, a reflexão do texto de Moscou na cultura e, principalmente, na literatura russa; A terceira parte é composta por traduções de alguns ensaios de escritores russos do século XIX nos quais foram descritas as relações controversas entre Moscou e São Petersburgo. A conclusão da presente pesquisa consiste na tese de que tanto o texto de Moscou como sua oposição ao texto de São Petersburgo formam-se como resultado da bipolaridade, própria não apenas da cultura russa, como de toda humanidade. / The thesis aims to study the Moscow city text, its comparative analysis and the dialogue with the St. Petersburg text, and the importance of both texts in the process of formation of Russian cultural identity. The survey methodology is based on the concept of urban text drawn up by prominent scholars of Tartu-Moscow Semiotic School, such as Yuri Lotamn, Vladimir Toporov, among others. The work presented is comprised of three parts: The first chapter presents the theoretical analysis of concepts of \"text\" and \"city\" and its application to Moscow; The second chapter represents a cultural-historical analysis of the formation process of the Moscow text since the founding of the city until the 21st century, the reflection of Moscow text in the culture and, mostly in the Russian literature; The third part is composed of translations of some Russian writers´ essays of the 19th century with a description of the controversial relations between Moscow and St. Petersburg. The conclusion of this research consists of the thesis that both Moscow text and its opposition to the St. Petersburg text are formed as a result of a proper bipolarity not only of the Russian culture, but of the whole mankind.

Dass die Textlinguistik nicht umhin kommt, ihre Beschreibungen und ihr Instrumentarium auch im Hinblick auf Texte zu überprüfen, die nicht von Muttersprachlern erzeugt worden sind, ist durch die wachsenden Bedürfnisse nach fremdsprachiger Kommunikation und die rapide steigende Zahl entsprechender kommunikativer Ereignisse bedingt. Der Fokus der vorliegenden Publikation liegt auf dem fremdsprachigen Text (vorzugsweise in seiner schriftlichen, monologischen Ausprägung) als sprachlicher Größe liegen. Bei der Analyse fremdsprachiger Texte werden neben textlinguistischen Fragen aber auch andere Gesichtspunkte berücksichtigt: soziolinguistische Aspekte, z.B. in Bezug auf die Kommunikationspartner, die kommunikativen Domänen, den sozial-historischen Rahmen der Kommunikation usw., psycholinguistische Hintergründe, z.B. im Hinblick auf die kognitive Verarbeitung fremdsprachlich-kommunikativer Anforderungen, literaturwissenschaftliche Bezüge, etwa hinsichtlich fremdsprachiger Einflüsse auf literarische Werke und ihre Rolle im literarischen Diskurs, translationswissenschaftliche Gesichtspunkte im Hinblick auf die Adaptation und Bewahrung des Fremden im übersetzten Text sowie pädagogisch-didaktische Fragen im Rahmen des gesteuerten Fremdsprachenerwerbs. / Owing to the growing necessity for foreign language interaction and the rapidly increasing number of associated communicative events, it is essential for text linguistics also to test its categories and methodology on texts which have not been produced by native speakers. The focus of the present publication lies on the foreign language text (primarily on its written, monological form) which is conceived as a linguistic entity. However, alongside textlinguistic issues the analysis of foreign language texts also entails a consideration of the following parameters: sociolinguistic aspects, e.g. with regard to the communicative partners, communicative domains, social-historic framework of the communicative situation etc.; psycholinguistic aspects, e.g. with respect to the cognitive processing of foreign language communication tasks; literary aspects, e.g. with regard to foreign language influences on literary works and their role in literary discourse; translatory aspects concerning the adaptation and preservation of foreign cultural elements in a translated text and pedagogical-didactic questions concerning foreign language teaching.

To date, hundreds of millions of mini-QWERTY keyboard equipped devices (miniaturized versions of a full desktop keyboard) have been sold. Accordingly, a large percentage of text messages originate from fixed-key, mini-QWERTY keyboard enabled mobile phones. Over a series of three longitudinal studies I quantify how quickly and accurately individuals can input text on mini-QWERTY keyboards. I evaluate performance in ideal laboratory conditions as well as in a variety of mobile contexts. My first study establishes baseline performance measures; my second study investigates the impact of limited visibility on text input performance; and my third study investigates the impact of mobility (sitting, standing, and walking) on text input performance. After approximately five hours of practice, participants achieved expertise typing almost 60 words per minute at almost 95% accuracy. Upon completion of these studies, I examine the types of errors that people make when typing on mini-QWERTY keyboards. Having discovered a common pattern in errors, I develop and refine an algorithm to automatically detect and correct errors in mini-QWERTY keyboard enabled text input. I both validate the algorithm through the analysis of pre-recorded typing data and then empirically evaluate the impacts of automatic error correction on live mini-QWERTY keyboard text input. Validating the algorithm over various datasets, I demonstrate the potential to correct approximately 25% of the total errors and correct up to 3% of the total keystrokes. Evaluating automatic error detection and correction on live typing results in successfully correcting 61% of the targeted errors committed by participants while increasing typing rates by almost two words per minute without introducing noticeable distraction.

In this dissertation, we introduce a novel text representation method mainly used for text classification purpose. The presented representation method is initially based on a variety of closeness relationships between pairs of words in text passages within the entire corpus. This representation is then used as the basis for our multi-level lightweight ontological representation method (TOR-FUSE), in which documents are represented based on their contexts and the goal of the learning task. The method is unlike the traditional representation methods, in which all the documents are represented solely based on the constituent words of the documents, and are totally isolated from the goal that they are represented for. We believe choosing the correct granularity of representation features is an important aspect of text classification. Interpreting data in a more general dimensional space, with fewer dimensions, can convey more discriminative knowledge and decrease the level of learning perplexity. The multi-level model allows data interpretation in a more conceptual space, rather than only containing scattered words occurring in texts. It aims to perform the extraction of the knowledge tailored for the classification task by automatic creation of a lightweight ontological hierarchy of representations. In the last step, we will train a tailored ensemble learner over a stack of representations at different conceptual granularities. The final result is a mapping and a weighting of the targeted concept of the original learning task, over a stack of representations and granular conceptual elements of its different levels (hierarchical mapping instead of linear mapping over a vector). Finally the entire algorithm is applied to a variety of general text classification tasks, and the performance is evaluated in comparison with well-known algorithms.

Textual information within digital media can be used in many areas such as, indexing and structuring of media databases, in the aid of visually impaired, translation of foreign signs and many more. This said, mainly text can be separated into two categories in digital media as, overlay-text and scene-text. In this thesis localization and recognition of video text regardless of its category in digital media is investigated. As a necessary first step, framework of a complete system is discussed. Next, a comparative analysis of feature vector and classification method pairs is presented. Furthermore, multi-part nature of text is exploited by proposing a novel Markov Random Field approach for the classification of text/non-text regions. Additionally, better localization of text is achieved by introducing bounding-box extraction method. And for the recognition of text regions, a handprint based Optical Character Recognition system is thoroughly investigated. During the investigation of text recognition, multi-hypothesis approach for the segmentation of background is proposed by incorporating k-Means clustering. Furthermore, a novel dictionary-based ranking mechanism is proposed for recognition spelling correction. And overall system is simulated on a challenging data set. Also, a through survey on scene-text localization and recognition is presented. Furthermore, challenges are identified and discussed by providing related work on them. Scene-text localization simulations on a public competition data set are also provided. Lastly, in order to improve recognition performance of scene-text on signs that are affected from perspective projection distortion, a rectification method is proposed and simulated.

This work aims the investigation of the several criminal suit court record text categories: types, genres and species (according to TRAVAGLIA, [2003]/2007) which are used in our society in the criminal court. Thus, it is necessary to contemplate the idea that the number of texts used in a criminal process is great and that each epoch lives a set of rules that are of its own, i.e, they do not abandon the past, they do not break up with the traditions, although they shape or discipline the human facts according to the demands of the moment. Therefore, our intention is to verify the existence of these several categories of texts, characterize them linguistically and verify in the genre sentence the reflexes, by means of linguistic marks, of those text genres which were used in criminal processes and motivated the judge to make up his decision and declare his sentence . This study was done according to the theoretical bases of the Text Linguistics and the reflections of the several text categories present in our daily routine at the criminal court, about which, lawyers manuals refer to, sometimes as criminal court documents, process or simply nominate them according to the legal terminology proposed in our criminal legal texts, which many times can lead to misunderstandings once they do not have the rigor of the juridical terminology, once the rigor of the law science depends substantially on the language which can not nominate different institutes with the same name neither can it nominate the same institutes with different names. Nevertheless, the legislator did not hold on any criteria and, being divorced of any scientific or systematic orientation, has opted for each law the empiric solutions. So, for the analyses of our text categories we have developed a qualitative study with an analytic-descriptive methodology that uses scraps of the collected corpus from 10 (ten) criminal processes collected for this investigation with the scope of analyzing and characterizing each of them with special attention to the sentence by means of the linguistic materiality of this corpus which is left behind as a clue to the motivation of the judge. Such analyses and characterizations are also based on the foundations of the Theory of the Speech Acts (AUSTIN, 1962), The Communicative Action Theory (HABERMAS, 1983, 2003), the theories about the argumentation, as well as the postulate of Discourse Communities (SWALES, 1990). / Este estudo tem como finalidade investigar as várias categorias de texto forenses criminal: tipos, gêneros e espécies (segundo TRAVAGLIA, [2003]/2007) que circulam em nossa sociedade. Destarte, é necessário contemplar a idéia de que grande é o número das categorias de texto que circulam em nossa sociedade e que cada época vive um complexo de regras que lhe são próprias, ou seja, não desprezam o passado, não rompem com as tradições, contudo, modelam ou disciplinam os fatos humanos, segundo as exigências do seu momento. Portanto pretende-se verificar a existência das várias categorias de texto presentes nos processos penais, caracterizá-las lingüisticamente e verificar no gênero sentença os reflexos, por meio das marcas lingüísticas, daqueles gêneros textuais apresentados no decorrer do processo criminal que motivam o julgador a proferir sua sentença. Norteia-se, este estudo, teoricamente pelos estudos da Lingüística Textual e pelas reflexões sobre as mais variadas categorias de texto presentes no nosso dia a dia forense, sobre as quais manuais de advogados ora se referem como peças processuais, ora como processo, ou simplesmente pela terminologia jurídica proposta na legislação, que muitas vezes leva ao erro, uma vez que não possuem um rigor da terminologia na linguagem jurídica. Haja vista, que o rigor da ciência jurídica depende substancialmente da linguagem, não devendo designar com um nome comum institutos diversos, nem institutos iguais com nomes diferentes. O legislador não abraçou, porém, nenhum critério e divorciado de qualquer preocupação científica ou sistemática, preferiu, em cada lei, as soluções meramente empíricas. Assim, para a análise das categorias de textos forenses criminais, desenvolve-se um estudo qualitativo, cuja metodologia analítico-descritiva faz recortes no corpus coletado em 10 (dez) processos criminais, analisa-se e caracteriza-se criteriosamente o gênero sentença por meio das marcas presentes na materialidade lingüística desse corpus. Tais análises e caracterizações norteiam-se, também, pelos pressupostos teóricos da Teoria dos Atos de Fala (AUSTIN, 1962), da teoria da Ação Comunicativa (HABERMAS, 1983,2003), as teorias sobre a argumentação, assim como pelo postulado teórico de Comunidades Discursivas (SWALES, 1990). / Mestre em Linguística

Orientadores: Jorge Stolfi, Neucimar Jerônimo Leite / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Computação / Made available in DSpace on 2018-08-20T03:12:07Z (GMT). No. of bitstreams: 1 Minetto_Rodrigo_D.pdf: 35894128 bytes, checksum: 8a0e453fba7e6a9a02fb17a52fdbf878 (MD5) Previous issue date: 2012 / Resumo: Nesta tese abordamos três problemas de visão computacional: (1) detecção e reconhecimento de objetos de texto planos em imagens de cenas reais; (2) rastreamento destes objetos de texto em vídeos digitais; e (3) o rastreamento de um objeto tridimensional rígido arbitrário com marcas conhecidas em um vídeo digital. Nós desenvolvemos, para cada um dos problemas, algoritmos inovadores, que são pelo menos tão precisos e robustos quanto outros algoritmos estado-da-arte. Especificamente, para reconhecimento de texto nós desenvolvemos (e validamos extensivamente) um novo descritor de imagem baseado em HOG especializado para escrita romana, que denominamos T-HOG, e mostramos sua contribuição como um filtro em um detector de texto (SNOOPERTEXT). Nós também melhoramos o algoritmo SNOOPERTEXT através do uso da técnica multiescala para tratar caracteres de tamanhos bastante variados e limitar a sensibilidade do algoritmo a vários artefatos. Para rastreamento de texto, nós descrevemos quatro estratégias básicas para combinar a detecção e o rastreamento de texto, e desenvolvemos também um rastreador específico baseado em filtro de partículas que explora o uso do reconhecedor T-HOG. Para o rastreamento de objetos rígidos, nós desenvolvemos um novo algoritmo preciso e robusto (AFFTRACK) que combina rastreamento de características por KLT com uma calibração de câmera melhorada. Nós testamos extensivamente nossos algoritmos com diversas bases de dados descritas na literatura. Nós também desenvolvemos algumas bases de dados (publicamente disponíveis) para a validação de algoritmos de detecção e rastreamento de texto e de rastreamento de objetos rígidos em vídeos / Abstract: In this thesis we address three computer vision problems: (1) the detection and recognition of flat text objects in images of real scenes; (2) the tracking of such text objects in a digital video; and (3) the tracking an arbitrary three-dimensional rigid object with known markings in a digital video. For each problem we developed innovative algorithms, which are at least as accurate and robust as other state-of-the-art algorithms. Specifically, for text classification we developed (and extensively evaluated) a new HOG-based descriptor specialized for Roman script, which we call T-HOG, and showed its value as a post-filter for an existing text detector (SNOOPERTEXT). We also improved the SNOOPERTEXT algorithm by using the multi-scale technique to handle widely different letter sizes while limiting the sensitivity of the algorithm to various artifacts. For text tracking, we describe four basic ways of combining a text detector and a text tracker, and we developed a specific tracker based on a particle-filter which exploits the T-HOG recognizer. For rigid object tracking we developed a new accurate and robust algorithm (AFFTRACK) that combines the KLT feature tracker with an improved camera calibration procedure. We extensively tested our algorithms on several benchmarks well-known in the literature. We also created benchmarks (publicly available) for the evaluation of text detection and tracking and rigid object tracking algorithms / Doutorado / Ciência da Computação / Doutor em Ciência da Computação

The subject of this thesis is the translation of wordplays and cultural words in a German tourist brochure into Swedish. The overall aim was to examine which strategies were used to translate them and how the skopos, i.e. the purpose, and the text type affected these choices. Translating wordplays is known to be notoriously difficult, and this is also brought to light in this thesis. Even though 55% of the wordplays could be translated into a wordplay, 32% could not. In a few cases the strategy wordplay à related rhetorical device (for example rhyme, metaphor or irony) was used in order to retain the vocative purpose of the wordplay. A stronger way to ensure and preserve the vocative function in the target text was by using the strategy non wordplay à wordplay, a so-called compensatory strategy – a strategy sometimes deemed necessary. Cultural words are concepts that demand a certain amount of knowledge of the reader. The aim was to examine if these concepts, such as material culture (for example food and towns) and ecology, should be translated by source-text oriented, or rather by target-text oriented translation strategies. The results show that both strategies are indispensable in order to fulfil the skopos of a tourist brochure. However, the target-text oriented strategies, such as generalization and substitution, dominated with 52%. The source-text oriented strategies were used in only 38% of the cases. In some cases the cultural word was translated by an official equivalent, i.e. neither a source- nor target-text oriented strategy.

This dissertation constitutes an illustration of how writing instruction in elementary school, year three, is conducted. At this stage of schooling, most pupils are assumed to have acquired basic reading and writing skills, and from now on, and increasingly over the years that follow, they are expected to read and write longer texts within different genres and subjects. The aim of this thesis is to study writing lessons that were conducted under the framework of four writing projects in a classroom, and the student-written texts that that resulted thereof. The writing pro-jects are characterised by focus on similar text types, which in my mate-rial includes "the fairy tale", "a letter to the editor", "instructions" and a "factual text". The focus of the study is on the relationship between the learning support offered to the pupils in the classroom and the na-ture of texts that the pupils then write. The research is inspired by ethnographic methodology, and is based on material consisting of field notes, video recordings and student texts. The theoretical framework assumes a socio-cultural view of learning and a dialogical view of text and writing. The teaching practices are studied on the basis of how they are built up by different chains of activities (reading, conversation and writing). Particular attention is paid to which text dimensions are addressed in classroom conversations: content, form or function. The pupils’ texts are analysed on the basis of their macro structure, and the analysis builds on the concept of "text activity". On an overall level, the results point to writing being a social activity, which is also closely interconnected to reading, and above all dialogue and conversation. The writing instruction offered to the students is also characterized by a broad view of what literacy is about. The study points out the value of versatile learning support, where the function, form and content of texts, in relation to the learning goals, are made explicit in the teaching. An important result is that the functional dimension of writing, in particular, favours writing development. On a more general level, the study raises the question of which literacy skills can and should be pro-moted by writing education in the early primary school years.

In this dissertation, we introduce a novel text representation method mainly used for text classification purpose. The presented representation method is initially based on a variety of closeness relationships between pairs of words in text passages within the entire corpus. This representation is then used as the basis for our multi-level lightweight ontological representation method (TOR-FUSE), in which documents are represented based on their contexts and the goal of the learning task. The method is unlike the traditional representation methods, in which all the documents are represented solely based on the constituent words of the documents, and are totally isolated from the goal that they are represented for. We believe choosing the correct granularity of representation features is an important aspect of text classification. Interpreting data in a more general dimensional space, with fewer dimensions, can convey more discriminative knowledge and decrease the level of learning perplexity. The multi-level model allows data interpretation in a more conceptual space, rather than only containing scattered words occurring in texts. It aims to perform the extraction of the knowledge tailored for the classification task by automatic creation of a lightweight ontological hierarchy of representations. In the last step, we will train a tailored ensemble learner over a stack of representations at different conceptual granularities. The final result is a mapping and a weighting of the targeted concept of the original learning task, over a stack of representations and granular conceptual elements of its different levels (hierarchical mapping instead of linear mapping over a vector). Finally the entire algorithm is applied to a variety of general text classification tasks, and the performance is evaluated in comparison with well-known algorithms.