Improving face recognition with multispectral fusion and support vector machines

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Composição de dicionários visuais utilizando agrupamento de dados por Florestas de Caminhos Ótimos

Pós-graduação em Ciência da Computação - IBILCE

Percepção viso-motora de escolares com Transtorno do Déficit de Atenção com Hiperatividade

PURPOSE: The aim of this study was to characterize and to compare the visual-motor perception of students with Attention Deficit with Hyperactivity Disorder (ADHD) with students with good academic performance. METHODS: Forty students from 2nd to 5th grades of an elementary public school, male gender (100%), aged between 7 and 10 years and 8 months old participated, divided into: GI (20 students with ADHD) and GII (20 students with good academic performance), paired according to age, schooling and gender with GI. The students were submitted to Developmental Test of Visual Perception (DTVP-2). RESULTS: The students of GI presented low performance in spatial position and visual closure (reduced motor) and inferior age equivalent in reduced motor perception, when compared to GII. CONCLUSION: The difficulties in visual-motor perception presented by students of GI cannot be attributed to a primary deficit, but to a secondary phenomenon of inattention that interferes directly in their visual-motor performance.

Registro global de nuvens de pontos RGB-D em tempo real usando fluxo óptico e marcadores

Registration of point clouds captured by depth sensors is an important task in 3D reconstruction applications based on computer vision. In many applications with strict performance requirements, the registration should be executed not only with precision, but also in the same frequency as data is acquired by the sensor. This thesis proposes theuse of the pyramidal sparse optical flow algorithm to incrementally register point clouds captured by RGB-D sensors (e.g. Microsoft Kinect) in real time. The accumulated errorinherent to the process is posteriorly minimized by utilizing a marker and pose graph optimization. Experimental results gathered by processing several RGB-D datasets validatethe system proposed by this thesis in visual odometry and simultaneous localization and mapping (SLAM) applications.

A contribution to semantic description of images and videos: an application of soft biometrics

Humans have a high ability to extract visual data information acquired by sight. Trought a learning process, which starts at birth and continues throughout life, image interpretation becomes almost instinctively. At a glance, one can easily describe a scene with reasonable precision, naming its main components. Usually, this is done by extracting low-level features such as edges, shapes and textures, and associanting them to high level meanings. In this way, a semantic description of the scene is done. An example of this, is the human capacity to recognize and describe other people physical and behavioral characteristics, or biometrics. Soft-biometrics also represents inherent characteristics of human body and behaviour, but do not allow unique person identification. Computer vision area aims to develop methods capable of performing visual interpretation with performance similar to humans. This thesis aims to propose computer vison methods which allows high level information extraction from images in the form of soft biometrics. This problem is approached in two ways, unsupervised and supervised learning methods. The first seeks to group images via an automatic feature extraction learning , using both convolution techniques, evolutionary computing and clustering. In this approach employed images contains faces and people. Second approach employs convolutional neural networks, which have the ability to operate on raw images, learning both feature extraction and classification processes. Here, images are classified according to gender and clothes, divided into upper and lower parts of human body. First approach, when tested with different image datasets obtained an accuracy of approximately 80% for faces and non-faces and 70% for people and non-person. The second tested using images and videos, obtained an accuracy of about 70% for gender, 80% to the upper clothes and 90% to lower clothes. The results of these case studies, show that proposed methods are promising, allowing the realization of automatic high level information image annotation. This opens possibilities for development of applications in diverse areas such as content-based image and video search and automatica video survaillance, reducing human effort in the task of manual annotation and monitoring.

Detecção de região saliente em imagens usando dissimilaridade de cor e amostragem por pixeis aleatórios

The purpose of this work is to demonstrate and to assess a simple algorithm for automatic estimation of the most salient region in an image, that have possible application in computer vision. The algorithm uses the connection between color dissimilarities in the image and the image’s most salient region. The algorithm also avoids using image priors. Pixel dissimilarity is an informal function of the distance of a specific pixel’s color to other pixels’ colors in an image. We examine the relation between pixel color dissimilarity and salient region detection on the MSRA1K image dataset. We propose a simple algorithm for salient region detection through random pixel color dissimilarity. We define dissimilarity by accumulating the distance between each pixel and a sample of n other random pixels, in the CIELAB color space. An important result is that random dissimilarity between each pixel and just another pixel (n = 1) is enough to create adequate saliency maps when combined with median filter, with competitive average performance if compared with other related methods in the saliency detection research field. The assessment was performed by means of precision-recall curves. This idea is inspired on the human attention mechanism that is able to choose few specific regions to focus on, a biological system that the computer vision community aims to emulate. We also review some of the history on this topic of selective attention.

Intelligent system for interaction with virtual characters based on volumetric sensors

Dissertação de Mestrado, Engenharia Elétrica e Eletrónica, Instituto Superior de Engenharia, Universidade do Algarve, 2015

Média-arte digital : arte na era do artefacto digital/computacional

Vivemos atualmente a denominada revolução tecnológica cujos contornos ou mesmo linhas futuras de desenvolvimento nos são ainda em larga medida desconhecidos fruto sobretudo da sua acelerada evolução que não cessa de nos surpreender com novos e imprevisíveis resultados. O computador pessoal, a rede Web, ou mesmo a massificação dos dispositivos móveis de informação e comunicação que hoje assistimos no mundo inteiro, são bons exemplos de marcos evolutivos da revolução tecnológica, em grande medida impensáveis há apenas uma ou duas gerações.

Navegação, reconhecimento de gestos e controle de interface no sistema ITV utilizando o dispositivo kinect

Esta dissertação apresenta a implementação de navegação no ambiente virtual, reconhecimento de gestos e controle de interface, feitos através do dispositivo Kinect, no Sistema ITV: um sistema de treinamento de operadores e mantenedores de usinas hidrelétricas e subestações elétricas. São mostrados, também, determinados aperfeiçoamentos recentes, como conversão em vídeo, telas de alarmes sonoros e visuais, ambientação sonora em três dimensões e narração do processo. Além da apresentação do Sistema ITV, são expostos o dispositivo Kinect e o algoritmo utilizado na comparação dos padrões de movimento, o DTW. Em seguida, são abordados em detalhes o projeto e a implementação da navegação, do reconhecimento de gestos e do controle de interface. Como estudo de caso, é exibida uma Instrução Técnica Virtual (ITV), elaborada especialmente para testar e avaliar a nova interface proposta. Posteriormente, são apresentados os resultados, considerados satisfatórios, obtidos através da análise de questionários qualitativos aplicados a estudantes da Universidade Federal do Pará. Por fim, são realizadas as considerações referentes a este trabalho e expostas idéias de trabalhos futuros.

Navegação cooperativa de um robô humanóide e um robô com rodas usando informação visual

This work presents a cooperative navigation systemof a humanoid robot and a wheeled robot using visual information, aiming to navigate the non-instrumented humanoid robot using information obtained from the instrumented wheeled robot. Despite the humanoid not having sensors to its navigation, it can be remotely controlled by infra-red signals. Thus, the wheeled robot can control the humanoid positioning itself behind him and, through visual information, find it and navigate it. The location of the wheeled robot is obtained merging information from odometers and from landmarks detection, using the Extended Kalman Filter. The marks are visually detected, and their features are extracted by image processing. Parameters obtained by image processing are directly used in the Extended Kalman Filter. Thus, while the wheeled robot locates and navigates the humanoid, it also simultaneously calculates its own location and maps the environment (SLAM). The navigation is done through heuristic algorithms based on errors between the actual and desired pose for each robot. The main contribution of this work was the implementation of a cooperative navigation system for two robots based on visual information, which can be extended to other robotic applications, as the ability to control robots without interfering on its hardware, or attaching communication devices

Orientação espacial em adultos com deficiência visual: efeitos de um treinamento de navegação

O objetivo deste estudo foi demonstrar se um programa de navegação pode ajudar indivíduos com deficiência visual a melhorar a acurácia na orientação dinâmica. Nove participantes com deficiência visual retornaram a um ponto de partida após percorrer rotas em linha reta e triangular. Pré e pós-avaliações foram feitas entre um período de 4 meses, durante o qual o treinamento com navegação foi realizado. Entre pré e pós-teste, erros relativos de desvios angulares (ERDA) foram diferentes apenas na tarefa em linha reta. O valor de ERDA foi maior na tarefa em linha reta possivelmente por causa da magnitude do giro inicial antes de retornar ao ponto de partida (i.e., 180º) em contraste com a tarefa triângulo (i.e., 45º). Conclui-se que, em tarefas de orientação, os erros no desvio angular dependem da amplitude do giro inicial ao retornar para o ponto de partida. Ainda, a acurácia na manutenção da direção é influenciada por um treinamento específico com navegação.

O papel dos sistemas visual, vestibular, somatosensorial e auditivo para o controle postural

The maintenance of a given body orientation is obtained by the complex relation between sensory information and muscle activity. Therefore, this study purpose was to review the role of visual, somatosensory, vestibular and auditory information in the maintenance and control of the posture. Method. a search by papers for the last 24 years was done in the PubMed and CAPES databases. The following keywords were used: postural control, sensory information, vestibular system, visual system, somatosensory system, auditory system and haptic system. Results. the influence of each sensory system and its integration were analyzed for the maintenance and control of the posture. Conclusion. the literature showed that there is information redundancy provided by sensory channels. Thus, the central nervous system chooses the main source for the posture control.

O Uso da informação visual global e local no controle da postura e do timing durante a tarefa de agarrar um objeto em aproximação

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Efeitos da manipulação do estímulo visual no controle postural nas faixas etárias de 4 a 14 anos de idade

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)