Scene Classification Using a Hybrid Generative/Discriminative Approach

We investigate whether dimensionality reduction using a latent generative model is beneficial for the task of weakly supervised scene classification. In detail, we are given a set of labeled images of scenes (for example, coast, forest, city, river, etc.), and our objective is to classify a new image into one of these categories. Our approach consists of first discovering latent ";topics"; using probabilistic Latent Semantic Analysis (pLSA), a generative model from the statistical text literature here applied to a bag of visual words representation for each image, and subsequently, training a multiway classifier on the topic distribution vector for each image. We compare this approach to that of representing each image by a bag of visual words vector directly and training a multiway classifier on these vectors. To this end, we introduce a novel vocabulary using dense color SIFT descriptors and then investigate the classification performance under changes in the size of the visual vocabulary, the number of latent topics learned, and the type of discriminative classifier used (k-nearest neighbor or SVM). We achieve superior classification performance to recent publications that have used a bag of visual word representation, in all cases, using the authors' own data sets and testing protocols. We also investigate the gain in adding spatial information. We show applications to image retrieval with relevance feedback and to scene classification in videos

Neural mechanisms of visual object priming: evidence for perceptual and semantic distinctions in fusiform cortex

Previous functional imaging studies have shown that facilitated processing of a visual object on repeated, relative to initial, presentation (i.e., repetition priming) is associated with reductions in neural activity in multiple regions, including fusiforin/lateral occipital cortex. Moreover, activity reductions have been found, at diminished levels, when a different exemplar of an object is presented on repetition. In one previous study, the magnitude of diminished priming across exemplars was greater in the right relative to the left fusiform, suggesting greater exemplar specificity in the right. Another previous study, however, observed fusiform lateralization modulated by object viewpoint, but not object exemplar. The present fMRI study sought to determine whether the result of differential fusiform responses for perceptually different exemplars could be replicated. Furthermore, the role of the left fusiform cortex in object recognition was investigated via the inclusion of a lexical/semantic manipulation. Right fusiform cortex showed a significantly greater effect of exemplar change than left fusiform, replicating the previous result of exemplar-specific fusiform lateralization. Right fusiform and lateral occipital cortex were not differentially engaged by the lexical/semantic manipulation, suggesting that their role in visual object recognition is predominantly in the. C visual discrimination of specific objects. Activation in left fusiform cortex, but not left lateral occipital cortex, was modulated by both exemplar change and lexical/semantic manipulation, with further analysis suggesting a posterior-to-anterior progression between regions involved in processing visuoperceptual and lexical/semantic information about objects. The results are consistent with the view that the right fusiform plays a greater role in processing specific visual form information about objects, whereas the left fusiform is also involved in lexical/semantic processing. (C) 2003 Elsevier Science (USA). All rights reserved.

Variável colativa por complexidade, seus efeitos no campo visual monocular

Os resultados das análises feitas com estes dados indicaram diferenças significativas no aumento da amplitude do plano meridiano horizontal nasal do campo visual monocular, medidas em unidades angulares. As diferenças foram interpretadas como indicativas da influência dos três diferentes níveis de complexidade dos estímulos visuais. Concluiu-se, portanto, que a variável colativa por complexidade influi no ato perceptual do reconhecimento visual.

Hybrid visual-driven decision support system in video monitor manufacturing

An intelligent system that emulates human decision behaviour based on visual data acquisition is proposed. The approach is useful in applications where images are used to supply information to specialists who will choose suitable actions. An artificial neural classifier aids a fuzzy decision support system to deal with uncertainty and imprecision present in available information. Advantages of both techniques are exploited complementarily. As an example, this method was applied in automatic focus checking and adjustment in video monitor manufacturing. Copyright © 2005 IFAC.

Image categorization through optimum path forest and visual words

Different from the first attempts to solve the image categorization problem (often based on global features), recently, several researchers have been tackling this research branch through a new vantage point - using features around locally invariant interest points and visual dictionaries. Although several advances have been done in the visual dictionaries literature in the past few years, a problem we still need to cope with is calculation of the number of representative words in the dictionary. Therefore, in this paper we introduce a new solution for automatically finding the number of visual words in an N-Way image categorization problem by means of supervised pattern classification based on optimum-path forest. © 2011 IEEE.

Automatic visual dictionary generation through Optimum-Path Forest clustering

Image categorization by means of bag of visual words has received increasing attention by the image processing and vision communities in the last years. In these approaches, each image is represented by invariant points of interest which are mapped to a Hilbert Space representing a visual dictionary which aims at comprising the most discriminative features in a set of images. Notwithstanding, the main problem of such approaches is to find a compact and representative dictionary. Finding such representative dictionary automatically with no user intervention is an even more difficult task. In this paper, we propose a method to automatically find such dictionary by employing a recent developed graph-based clustering algorithm called Optimum-Path Forest, which does not make any assumption about the visual dictionary's size and is more efficient and effective than the state-of-the-art techniques used for dictionary generation. © 2012 IEEE.

Automatic visual dictionary generation through optimum-path forest clusteringautomatic visual dictionary generation through optimum-path forest clustering

Image categorization by means of bag of visual words has received increasing attention by the image processing and vision communities in the last years. In these approaches, each image is represented by invariant points of interest which are mapped to a Hilbert Space representing a visual dictionary which aims at comprising the most discriminative features in a set of images. Notwithstanding, the main problem of such approaches is to find a compact and representative dictionary. Finding such representative dictionary automatically with no user intervention is an even more difficult task. In this paper, we propose a method to automatically find such dictionary by employing a recent developed graph-based clustering algorithm called Optimum-Path Forest, which does not make any assumption about the visual dictionary's size and is more efficient and effective than the state-of-the-art techniques used for dictionary generation.

Ecóico e nomeação de figuras em crianças com deficiência auditiva pré-lingual com implante coclear

The aim of this work was to verify the effect of teaching the echoic behavior over the pictures naming in four children between eight and nine years old with prelingual hearing impaired, users of cochlear implants. The design adopted was: (a) pre-training that taught the matching-to-sample task; (b) pre-tests that selected three words to teach; (c) teaching of auditory-visual conditional relations; (d) naming pos-test; (e) the teaching of echoic with orofacial clues and, (f) the second naming pos-test. In the pre-test all participants achieved smaller percentage of correct on naming (60%-80%) and echoic (20%-50%) when compared to percentages word recognition (86%-93%). All participants learned the auditory-visual relations. The improvement on naming test occurred after auditory training select based for two participants; for other two participants the improvement on naming test occurred just after the training of echoic. Analysis of data showed that the listening and speaking performances are independent in their establishment and require specific conditions of teaching; in the case of this study, even though the result is not generalized to all participants, the highest correspondence into point to point naming was obtained following the teaching of echoic.

About the role of visual field defects in pure alexia

Pure alexia is an acquired reading disorder characterized by a disproportionate prolongation of reading time as a function of word length. Although the vast majority of cases reported in the literature show a right-sided visual defect, little is known about the contribution of this low-level visual impairment to their reading difficulties. The present study was aimed at investigating this issue by comparing eye movement patterns during text reading in six patients with pure alexia with those of six patients with hemianopic dyslexia showing similar right-sided visual field defects. We found that the role of the field defect in the reading difficulties of pure alexics was highly deficit-specific. While the amplitude of rightward saccades during text reading seems largely determined by the restricted visual field, other visuo-motor impairments-particularly the pronounced increases in fixation frequency and viewing time as a function of word length-may have little to do with their visual field defect. In addition, subtracting the lesions of the hemianopic dyslexics from those found in pure alexics revealed the largest group differences in posterior parts of the left fusiform gyrus, occipito-temporal sulcus and inferior temporal gyrus. These regions included the coordinate assigned to the centre of the visual word form area in healthy adults, which provides further evidence for a relation between pure alexia and a damaged visual word form area. Finally, we propose a list of three criteria that may improve the differential diagnosis of pure alexia and allow appropriate therapy recommendations.

Graph-based Semi-Supervised Learning in Acoustic Modeling for Automatic Speech Recognition

Thesis (Ph.D.)--University of Washington, 2016-06

Psychophysical investigations of human peripheral vision

This thesis investigates various aspects of peripheral vision, which is known not to be as acute as vision at the point of fixation. Differences between foveal and peripheral vision are generally thought to be of a quantitative rather than a qualitative nature. However, the rate of decline in sensitivity between foveal and peripheral vision is known to be task dependent and the mechanisms underlying the differences are not yet well understood. Several experiments described here have employed a psychophysical technique referred to as 'spatial scaling'. Thresholds are determined at several eccentricities for ranges of stimuli which are magnified versions of one another. Using this methodology a parameter called the E2 value is determined, which defines the eccentricity at which stimulus size must double in order to maintain performance equivalent to that at the fovea. Experiments of this type have evaluated the eccentricity dependencies of detection tasks (kinetic and static presentation of a differential light stimulus), resolution tasks (bar orientation discrimination in the presence of flanking stimuli, word recognition and reading performance), and relative localisation tasks (curvature detection and discrimination). Most tasks could be made equal across the visual field by appropriate magnification. E2 values are found to vary widely dependent on the task, and possible reasons for such variations are discussed. The dependence of positional acuity thresholds on stimulus eccentricity, separation and spatial scale parameters is also examined. The relevance of each factor in producing 'Weber's law' for position can be determined from the results.

Aggregating Local Descriptors for Epigraphs Recognition

In this paper, we consider the task of recognizing epigraphs in images such as photos taken using mobile devices. Given a set of 17,155 photos related to 14,560 epigraphs, we used a k-NearestNeighbor approach in order to perform the recognition. The contribution of this work is in evaluating state-of-the-art visual object recognition techniques in this specific context. The experimental results conducted show that Vector of Locally Aggregated Descriptors obtained aggregating SIFT descriptors is the best choice for this task.

Examining Cognitive Processing in Children with an Arithmetic Disability

Currently there is no consensus as to the specific cognitive impairments that characterize mathematical disabilities (MD) or specific subtypes such as an arithmetic disability (AD). The present study sought to address this concern by examining cognitive processes that might undergird AD in children. The present study utilized archival data to conduct two investigations. The first investigation examined the executive functioning and working memory of children with AD. An age-matched achievement-matched design was employed to explore whether children with AD exhibit developmental lags or deficits in these cognitive domains. While children with AD did not exhibit impairments in verbal working memory or colour word inhibition, they did demonstrate impairments in shifting attention, visual-spatial working memory, and quantity inhibition. As children with AD did not perform more poorly than their younger achievement-matched peers on any of these tasks, impairments in specific areas of executive functioning and working memory appeared to reflect a developmental lag rather than a cognitive deficit. The second study examined the phonological processing performance of children with AD compared to children with comorbid disabilities in arithmetic and word recognition (AD/WRD) and to typically achieving (TA) children. Results indicated that, while children with AD did demonstrate impairments on all isolated naming speed tasks, trail making digits, and memory for digits, they did not demonstrate impairments on measures of phonological awareness, nonword repetition, serial processing speed, or serial naming speed. In contrast, children with AD/WRD demonstrated impairments on measures of phonological awareness, phonological short-term memory, isolated naming speed, serial processing speed, and the alphabet a-z task. Overall, results suggested that phonological processing impairments are more prominent in children with a WRD than children with an AD. Together, these studies further our understanding of the nature of the cognitive processes that underlie AD by focusing upon rarely used methods (i.e., age-matched achievement-matched design) and under-examined cognitive domains (i.e., phonological processing).

LEARNING A COMPLEX MOTOR SKILL FROM VIDEO AND POINT-LIGHT DEMONSTRATIONS

The aim of this Study was to compare the learning process of a highly complex ballet skill following demonstrations of point light and video models 16 participants divided into point light and video groups (ns = 8) performed 160 trials of a pirouette equally distributed in blocks of 20 trials alternating periods of demonstration and practice with a retention test a day later Measures of head and trunk oscillation coordination d1 parity from the model and movement time difference showed similarities between video and point light groups ballet experts evaluations indicated superiority of performance in the video over the point light group Results are discussed in terms of the task requirements of dissociation between head and trunk rotations focusing on the hypothesis of sufficiency and higher relevance of information contained in biological motion models applied to learning of complex motor skills