Thesis (M.S.)--University of California, Santa Cruz, 1988. / Typescript. Includes bibliographical references.

Thesis (M.S.)--Oregon State University, 2007. / Printout. Includes bibliographical references (leaves 44-45). Also available on the World Wide Web.

Thesis (M.S.)--Florida State University, 2006. / Advisor: Renatus Mussa, Florida State University, College of Engineering, Dept. of Civil and Environmental Engineering Title and description from dissertation home page (viewed Sept. 19, 2006). Document formatted into pages; contains xi, 202 pages. Includes bibliographical references.

Thesis (Ph.D.)--Ohio University, August, 2004. / Title from PDF t.p. Includes bibliographical references (p. 100-105)

Thesis (Ph. D.)--University of Texas at Austin, 2003. / Vita. Includes bibliographical references (p.137-141). Also available online,

Made available in DSpace on 2014-06-11T19:24:01Z (GMT). No. of bitstreams: 0 Previous issue date: 2013-05-15Bitstream added on 2014-06-13T19:30:28Z : No. of bitstreams: 1 iwashita_as_me_sjrp.pdf: 872507 bytes, checksum: 3a65e582b2dc78e6845ccfe3ddc8aa3f (MD5) / Técnicas de reconhecimento de padrões têm como principal objetivo classificar um conjunto de amostras baseadas em um conhecimento a prioriou em alguma informação estatística obtida dessas amostras. Tal processo de aprendizado é a fase de maior consumo de tempo na grande maioria das técnicas de reconhecimento de padrõe. O problema ainda pode piorar em ferramentas de classificação interativas, nas quais o usuário é solicitado a rotular amostras que serão utilizadas para o treinamento, e após a classificação, os resultados podem ser refina-dos através de mais amostras rotuladas manualmente. Esta situação pode ser inaceitável para grandes bases de dados. Dado que muitos trabalhos tem sido orientados à implementação de vários algoritmos de reconhecimento de padrôes em ambiente General Purpose Graphics Processing Unit- GPGPU, o presente estudo objetivou a implementação da etapa de treinamento do classificador Floresta de Caminhos Ótimos em Compute Unified Device Architecture- CUDA visando aumentar a sua eficiência. Foi implementada uma otimização, do referido classificador utilizando os métodos tradicionais, ou seja, na Central Processing Unit- CPU, e demonstrou uma fase de treinamento cerca de duas vezes mais rápida que a versão original. A otimização do classificador em CUDA também demonstrou uma fase de treinamento mais rápida que a versão original / Pattern recognition techniques have as main objective to classify a set of samples ba-sed on a priori knowledge or statistical information obtained by these samples. This learning process is the most time-consuming phase in most pattern recognition techniques. The problem may become worse in interactive classification tools, in which the user is asked to label the samples that will be used for training, and after the classification, the results can be refined through more samples manually labeled. However, this may be unacceptable for large databa-ses. Since many studies have been oriented to the implementation of various pattern recognition algorithms on General Purpose Graphics Processing Unit - GPGPU environment, this study ai-med the implementation of the training stage of the Optimum-Path Forest classifier in Compute Unified Device Architecture - CUDA in order to increase its efficiency. We have implemented an optimization of that classifier using the traditional methods, i.e., on the Central Processing Unit - CPU, and it has demonstrated a training phase about two times faster than the original version. The classifier optimization in CUDA has also shown a training phase faster than the original version

Orientador: João Paulo Papa / Coorientador: Alexandro José Baldassin / Banca: Antonio Carlos Sementille / Banca: Alexandre Luís Magalhães Levada / Resumo: Técnicas de reconhecimento de padrões têm como principal objetivo classificar um conjunto de amostras baseadas em um conhecimento a prioriou em alguma informação estatística obtida dessas amostras. Tal processo de aprendizado é a fase de maior consumo de tempo na grande maioria das técnicas de reconhecimento de padrõe. O problema ainda pode piorar em ferramentas de classificação interativas, nas quais o usuário é solicitado a rotular amostras que serão utilizadas para o treinamento, e após a classificação, os resultados podem ser refina-dos através de mais amostras rotuladas manualmente. Esta situação pode ser inaceitável para grandes bases de dados. Dado que muitos trabalhos tem sido orientados à implementação de vários algoritmos de reconhecimento de padrôes em ambiente General Purpose Graphics Processing Unit- GPGPU, o presente estudo objetivou a implementação da etapa de treinamento do classificador Floresta de Caminhos Ótimos em Compute Unified Device Architecture- CUDA visando aumentar a sua eficiência. Foi implementada uma otimização, do referido classificador utilizando os métodos tradicionais, ou seja, na Central Processing Unit- CPU, e demonstrou uma fase de treinamento cerca de duas vezes mais rápida que a versão original. A otimização do classificador em CUDA também demonstrou uma fase de treinamento mais rápida que a versão original / Abstract: Pattern recognition techniques have as main objective to classify a set of samples ba-sed on a priori knowledge or statistical information obtained by these samples. This learning process is the most time-consuming phase in most pattern recognition techniques. The problem may become worse in interactive classification tools, in which the user is asked to label the samples that will be used for training, and after the classification, the results can be refined through more samples manually labeled. However, this may be unacceptable for large databa-ses. Since many studies have been oriented to the implementation of various pattern recognition algorithms on General Purpose Graphics Processing Unit - GPGPU environment, this study ai-med the implementation of the training stage of the Optimum-Path Forest classifier in Compute Unified Device Architecture - CUDA in order to increase its efficiency. We have implemented an optimization of that classifier using the traditional methods, i.e., on the Central Processing Unit - CPU, and it has demonstrated a training phase about two times faster than the original version. The classifier optimization in CUDA has also shown a training phase faster than the original version / Mestre

Much of our interaction with the environment is physical. We use our bodies for nonverbal expression or to augment or emphasize verbal communication. In other cases we use our bodies to execute tasks such as walking or picking up an object. A human observer can easily recognise these activities. For example, it is the job of a security officer in a supermarket to observe people and check that articles are not stolen. If a person does steal, the security officer recognises the act and takes appropriate action. The problem addressed in this study is the automatic recognition of human gestures by means of video image analysis. For this purpose a computer-based system with similar recognition capabilities as a human observer is investigated. The system uses cameras that correspond to the eyes and algorithms that resemble abilities of the human visual system. Automatic gesture recognition is a complex problem and the focus here is to develop algorithms that will solve a subset of the problem. This involves the recognition of simple gestures such as walking and waving of arms. The approach taken in this dissertation is to represent body shape in camera images with a simple model called a bounding box. This model has the appearance of a rect¬angle that encapsulates the extremities of the human body and resembles the coarse structure of body shape. From a representation point of view, the model is an abstrac¬tion of body pose. A gesture consists of a sequence of poses. By employing pattern recognition techniques, a sequence of pose abstractions is recognised as a gesture. Various aspects of the bounding box model are explored in this study. Perception experiments are conducted to gain a conceptual understanding of the behaviour of the model. Other aspects include investigation of two- and three-dimensional spatial representations of the model with a neural network classifier as well as the model's temporal properties through the use of hidden Markov models. These aspects are tested using gesture recognition systems implemented for this purpose. The gesture vocabularies of these systems range from four to ten gestures, while recognition rates vary from 84.7% to 96.3%. / Dissertation (MEng (Computer Engineering))--University of Pretoria, 2006. / Electrical, Electronic and Computer Engineering / unrestricted

No description available.

No description available.