The time course of texture segregation was studied for three different measures of segregation (detection, localization and identification of an embedded texture region) under three different raster width conditions (26$ sp prime$, 43$ sp prime$ and 61$ sp prime$ of arc) by using a backward masking paradigm. The masking data were described with an exponential model the parameters of which represent rate of performance improvement and asymptotic performance level. The results indicate that: (1) information supporting localization accrues more rapidly than information supporting identification, (2) increasing element spacing had a more detrimental effect on identification than on localization, (3) under most spacing conditions, performance on detection fell between that for localization and identification. In conclusion, these three widely used texture segregation tasks cannot be considered equivalent measures of a single process. However, comparisons among their respective time courses may enable us to better characterize the mechanisms underlying the segregation process.

No description available.

Thesis (M.A.)--York University, 2003. Graduate Programme in Psychology. / Typescript. Includes bibliographical references (leaves 102-107). Also available on the Internet. MODE OF ACCESS via web browser by entering the following URL: http://wwwlib.umi.com/cr/yorku/fullcit?pMQ82940.

Thesis (M.S.)--Rochester Institute of Technology, 1994. / Typescript. Bibliography: leaves 57-58.

Abstract not available

Thesis (Ph. D.)--University of Missouri-Columbia, 2001. / Typescript. Vita. Includes bibliographical references (leaves 116-119). Also available on the Internet.

Thesis (Ph. D.)--University of Missouri-Columbia, 2001. / Typescript. Vita. Includes bibliographical references (leaves 116-119). Also available on the Internet.

This work has the goal of explore the triangulated mesh visualization problem in order that geometry and texture information could be computed adaptively according to the view position. As only the essential information from the mesh is going to persist, the main applications involved are linked with real-time mesh visualization which concerns a large number of scientific areas. To adapt the mesh geometry, its resolution is being locally changed through the insertion/removal of vertices, edges and faces, view-dependent considering the number of polygons to preserve smooth forms near to the original mesh. In the texture adaptation, the texture map association of the mesh is being adapted to the resolution changes minimizing or eliminating the texture distortion resulting from the collapse of vertices, edges and faces / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Este trabalho trata do problema de visualização de malhas triangulares de modo que a geometria e a textura das mesmas possam ser calculadas adaptativamente de acordo com a posição do orientador. Como apenas as informações necessárias de malha tendem a permanecer, as principais aplicações envolvidas estão relacionadas à visualização de malhas em tempo real, as quais têm sido de grande interesse de pesquisa em diversas áreas científicas. Na adaptação da geometria da malha, sua resolução vai sendo localmente alterada através de inserção/remoção de vértices, arestas e faces, de acordo com a posição do observador, de modo que o número de polígonos preserve uma forma visual suave e bem próxima da malha original. Na adaptação da textura, o mapeamento de textura associado `a malha vai sendo adaptado `as mudança as de sua resolução de modo a reduzir ou eliminar as distorções de textura resultantes do colapso de vértices, arestas e faces

Ces travaux de recherche ont pour ambition de contribuer à l'automatisation de l'inspection visuelle, dans le cadre du contrôle qualité de pièces métalliques à géométrie complexe. En soi, de nombreuses techniques d'optique, de numérisation, d'implémentation de rendu photo-réaliste, de classification d'images ou de données, et de reconnaissance de formes sont déjà fortement développées et appliquées chacune dans des domaines particuliers. Or, elles ne sont pas, ou rarement pour des cas particuliers, combinées pour obtenir une méthode complète de numérisation de l'apparence jusqu'à la reconnaissance, effective et perceptuelle, de l'objet et des anomalies esthétiques.Ces travaux ont profité des avancements des thèses précédentes sur la formalisation du contrôle qualité ainsi que sur un système agile de numérisation d'aspect de surface permettant la mise en évidence de toute la diversité d'anomalies esthétiques de surfaces. Ainsi, la contribution majeure réside dans l'adaptation des méthodes de traitement d'images à la structure formalisée du contrôle qualité, au format riche des données d'apparence et aux méthodes de classification pour réaliser la reconnaissance telle que le contrôleur humain.En ce sens, la thèse propose un décryptage des différentes méthodologies liées au contrôle qualité, au comportement du contrôleur humain, aux anomalies d'aspect de surface, aux managements et traitements de l'information visuelle, jusqu'à la combinaison de toutes ces contraintes pour obtenir un système de substitution partielle au contrôleur humain. L'objectif de la thèse, et du décryptage, est d'identifier et de réduire les sources de variabilité pour obtenir un meilleur contrôle qualité, notamment par l'automatisation intelligente et structurée de l'inspection visuelle. A partir d'un dispositif de vision par ordinateur choisi, la solution proposée consiste à analyser la texture visuelle. Celle est considérée en tant que signature globale de l'information d'apparence visuelle supérieure à une unique image contenant des textures images. L'analyse est effectuée avec des mécanismes de reconnaissance de formes et d'apprentissage machine pour établir la détection et l'évaluation automatiques d'anomalies d'aspect. / This research has the ambition to contribute to the automation of visual inspection, in the quality control of complex geometry metal parts. Firstly, many optical techniques, scanning, implementation of photorealistic rendering, classification of images or data, and pattern recognition are already highly developed and applied in each particular areas. But they are not, or rarely, in special cases, combined for a complete scanning method of appearance to the recognition, effective and perceptual, of object and aesthetic anomalies.This work benefited from the advancements of previous thesis on the formalization of quality control, as well as an agile system of surface appearance scanning to highlight the diversity of aesthetic anomalies surfaces. Thus, the major contribution lies in the adaptation of image processing methods to the formal structure of quality control, rich appearance data format and classification methods to achieve recognition as the human controller.In this sense, the thesis deciphers the different methodologies related to quality control, the human controller processes, surface appearance defects, the managements and processing of visual information, to the combination of all these constraints for a partial substitution system of the human controller. The aim of the thesis is to identify and reduce sources of variability to obtain better quality control, including through the intelligent and structured automation of visual inspection. From a selected computer vision device, the proposed solution is to analyze visual texture. This is regarded as a global signature of superior visual appearance information to a single image containing images textures. The analysis is performed with pattern recognition and machine learning mechanisms to develop automatic detection and evaluation of appearance defects.