published_or_final_version / abstract / toc / Computer Science and Information Systems / Doctoral / Doctor of Philosophy

Thesis (M.S.)--State University of New York at Binghamton, Department of Computer Science, Thomas J. Watson School of Engineering and Applied Science, 2007. / Includes bibliographical references.

Thesis (M.A.)--Baylor University, 2008. / Includes bibliographical references (p. 104-117)

Thesis (M.S.)--Rochester Institute of Technology, 1994. / Typescript. Includes bibliographical references (leaves 45-49).

Thesis (Ph. D.)--University of Hong Kong, 2005. / Title proper from title frame. Also available in printed format.

Thesis (M.Phil.) - University of Queensland, 2002. / Includes bibliography.

Thesis (Ph.D.)--University of Waikato, 2006. / Includes bibliographical references (p. 157-170)

Thesis (Ph. D.)--University of Washington, 2007. / Vita. Includes bibliographical references (leaves 129-139).

Emotions and sentiments play a crucial role in our everyday lives. They aid decision-making, learning, communication, and situation awareness in human-centric environments. Over the past two decades, researchers in artificial intelligence have been attempting to endow machines with cognitive capabilities to recognize, infer, interpret and express emotions and sentiments. All such efforts can be attributed to affective computing, an interdisciplinary field spanning computer science, psychology, social sciences and cognitive science. Sentiment analysis and emotion recognition has also become a new trend in social media, avidly helping users understand opinions being expressed on different platforms in the web. In this thesis, we focus on developing novel methods for text-based sentiment analysis. As an application of the developed methods, we employ them to improve multimodal polarity detection and emotion recognition. Specifically, we develop innovative text and visual-based sentiment-analysis engines and use them to improve the performance of multimodal sentiment analysis. We begin by discussing challenges involved in both text-based and multimodal sentiment analysis. Next, we present a number of novel techniques to address these challenges. In particular, in the context of concept-based sentiment analysis, a paradigm gaining increasing interest recently, it is important to identify concepts in text; accordingly, we design a syntaxbased concept-extraction engine. We then exploit the extracted concepts to develop conceptbased affective vector space which we term, EmoSenticSpace. We then use this for deep learning-based sentiment analysis, in combination with our novel linguistic pattern-based affective reasoning method termed sentiment flow. Finally, we integrate all our text-based techniques and combine them with a novel deep learning-based visual feature extractor for multimodal sentiment analysis and emotion recognition. Comparative experimental results using a range of benchmark datasets have demonstrated the effectiveness of the proposed approach.

Orientadores: Neucimar Jerônimo Leite, Hélio Pedrini / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Computação / Made available in DSpace on 2018-08-19T05:52:55Z (GMT). No. of bitstreams: 1 Valio_FelipeBraunger_M.pdf: 3505580 bytes, checksum: 3b20a046a5822011c617729904457d95 (MD5) Previous issue date: 2011 / Resumo: Detecção de textos em imagens é um problema que vem sendo estudado a várias décadas. Existem muitos trabalhos que estendem os métodos existentes para uso em análise de vídeos, entretanto, poucos deles criam ou adaptam abordagens que consideram características inerentes dos vídeos, como as informações temporais. Um problema particular dos vídeos, que será o foco deste trabalho, é o de detecção de legendas. Uma abordagem rápida para localizar quadros de vídeos que contenham legendas é proposta baseada em uma estrutura de dados especial denominada ritmo visual. O método é robusto à detecção de legendas com respeito ao alfabeto utilizado, ao estilo de fontes, à intensidade de cores e à orientação das legendas. Vários conjuntos de testes foram utilizados em nosso experimentos para demonstrar a efetividade do método / Abstract: Detection of text in images is a problem that has been studied for several decades. There are many works that extend the existing methods for use in video analysis, however, few of them create or adapt approaches that consider the inherent characteristics of video, such as temporal information. A particular problem of the videos, which will be the focus of this work, is the detection of subtitles. A fast method for locating video frames containing captions is proposed based on a special data structure called visual rhythm. The method is robust to the detection of legends with respect to the used alphabet, font style, color intensity and subtitle orientation. Several datasets were used in our experiments to demonstrate the effectiveness of the method / Mestrado / Ciência da Computação / Mestre em Ciência da Computação