A common method for finding an object's pose is the generalized Hough transform, which accumulates evidence for possible coordinate transformations in a parameter space and takes large clusters of similar transformations as evidence of a correct solution. We analyze this approach by deriving theoretical bounds on the set of transformations consistent with each data-model feature pairing, and by deriving bounds on the likelihood of false peaks in the parameter space, as a function of noise, occlusion, and tessellation effects. We argue that blithely applying such methods to complex recognition tasks is a risky proposition, as the probability of false positives can be very high.

Determining the identity and pose of occluded objects from noisy data is a critical step in interacting intelligently with an unstructured environment. Previous work has shown that local measurements of position and surface orientation may be used in a constrained search process to solve this problem, for the case of rigid objects, either two-dimensional or three-dimensional. This paper considers the more general problem of recognizing and locating objects that can vary in parameterized ways. We consider objects with rotational, translational, or scaling degrees of freedom, and objects that undergo stretching transformations. We show that the constrained search method can be extended to handle the recognition and localization of such generalized classes of object families.

Determining the identity and pose of occluded objects from noisy data is a critical part of a system's intelligent interaction with an unstructured environment. Previous work has shown that local measurements of the position and surface orientation of small patches of an object's surface may be used in a constrained search process to solve this problem for the case of rigid polygonal objects using two-dimensional sensory data, or rigid polyhedral objects using three-dimensional data. This note extends the recognition system to deal with the problem of recognizing and locating curved objects. The extension is done in two dimensions, and applies to the recognition of two-dimensional objects from two-dimensional data, or to the recognition of three-dimensional objects in stable positions from two- dimensional data.

Recognition and tracking of multiple objects in video sequences is one of the main challenges in computer vision that currently deserves a lot of attention from researchers. Almost all the reported approaches are very application-dependent and there is a lack of a general methodology for dynamic object recognition and tracking that can be instantiated in particular cases. In this thesis, the work is oriented towards the definition and development of such a methodology which integrates object recognition and tracking from a general perspective using a probabilistic framework called PIORT (probabilistic integrated object recognition and tracking framework). It include some modules for which a variety of techniques and methods can be applied. Some of them are well-known but other methods have been designed, implemented and tested during the development of this thesis.The first step in the proposed framework is a static recognition module that provides class probabilities for each pixel of the image from a set of local features. These probabilities are updated dynamically and supplied to a tracking decision module capable of handling full and partial occlusions. The two specific methods presented use RGB colour features and differ in the classifier implemented: one is a Bayesian method based on maximum likelihood and the other one is based on a neural network. The experimental results obtained have shown that, on one hand, the neural net based approach performs similarly and sometimes better than the Bayesian approach when they are integrated within the tracking framework. And on the other hand, our PIORT methods have achieved better results when compared to other published tracking methods. All these methods have been tested experimentally in several test video sequences taken with still and moving cameras and including full and partial occlusions of the tracked object in indoor and outdoor scenarios in a variety of cases with different levels of task complexity. This allowed the evaluation of the general methodology and the alternative methods that compose these modules.A Probabilistic Integrated Object Recognition and Tracking Framework for Video Sequences / El reconocimiento y seguimiento de múltiples objetos en secuencias de vídeo es uno de los principales desafíos en visión por ordenador que actualmente merece mucha atención de los investigadores. Casi todos los enfoques reportados son muy dependientes de la aplicación y hay carencia de una metodología general para el reconocimiento y seguimiento dinámico de objetos, que pueda ser instanciada en casos particulares. En esta tesis, el trabajo esta orientado hacia la definición y desarrollo de tal metodología, la cual integra reconocimiento y seguimiento de objetos desde una perspectiva general usando un marco probabilístico de trabajo llamado PIORT (Probabilistic Integrated Object Recognition and Tracking). Este incluye algunos módulos para los que se puede aplicar una variedad de técnicas y métodos. Algunos de ellos son bien conocidos, pero otros métodos han sido diseñados, implementados y probados durante el desarrollo de esta tesis.El primer paso en el marco de trabajo propuesto es un módulo estático de reconocimiento que provee probabilidades de clase para cada píxel de la imagen desde un conjunto de características locales. Estas probabilidades son actualizadas dinámicamente y suministradas a un modulo decisión de seguimiento capaz de manejar oclusiones parciales o totales. Se presenta dos métodos específicos usando características de color RGB pero diferentes en la implementación del clasificador: uno es un método Bayesiano basado en la máxima verosimilitud y el otro método está basado en una red neuronal. Los resultados experimentales obtenidos han mostrado que, por una parte, el enfoque basado en la red neuronal funciona similarmente y algunas veces mejor que el enfoque bayesiano cuando son integrados dentro del marco probabilístico de seguimiento. Por otra parte, nuestro método PIORT ha alcanzado mejores resultados comparando con otros métodos de seguimiento publicados. Todos estos métodos han sido probados experimentalmente en varias secuencias de vídeo tomadas con cámaras fijas y móviles incluyendo oclusiones parciales y totales del objeto a seguir, en ambientes interiores y exteriores, en diferentes tareas y niveles de complejidad. Esto ha permitido evaluar tanto la metodología general como los métodos alternativos que componen sus módulos.

In this work we address the problem of object recognition and localization from sparse range data. The method is based upon comparing the 7-D error surfaces of objects in various poses, which result from the registration error function between two convolved surfaces. The objects and their pose values are encoded by a small set of feature vectors extracted from the minima of the error surfaces. The problem of object recognition is thus reduced to comparing these feature vectors to find the corresponding error surfaces between the runtime data and a preprocessed database. Specifically, we present a new approach to the problems of pose determination, object recognition and object class recognition. The algorithm has been implemented and tested on both simulated and real data. The experimental results demonstrate the technique to be both effective and efficient, executing at 122 frames per second on standard hardware and with recognition rates exceeding 97% for a database of 60 objects. The performance of the proposed potential well space embedding (PWSE) approach on large size databases was also evaluated on the Princeton Shape Bench- mark containing 1,814 objects. In experiments of object class recognition with the Princeton Shape Benchmark, PWSE is able to provide better classification rates than the previous methods in terms of nearest neighbour classification. In addition, PWSE is shown to (i) operate with very sparse data, e.g., comprising only hundreds of points per image, and (ii) is robust to measurement error and outliers. / Thesis (Ph.D, Electrical & Computer Engineering) -- Queen's University, 2010-01-24 23:07:30.108

No description available.

No description available.

We present a component-based approach for recognizing objects under large pose changes. From a set of training images of a given object we extract a large number of components which are clustered based on the similarity of their image features and their locations within the object image. The cluster centers build an initial set of component templates from which we select a subset for the final recognizer. In experiments we evaluate different sizes and types of components and three standard techniques for component selection. The component classifiers are finally compared to global classifiers on a database of four objects.

We present a component-based approach for recognizing objectsunder large pose changes. From a set of training images of a givenobject we extract a large number of components which are clusteredbased on the similarity of their image features and their locations withinthe object image. The cluster centers build an initial set of componenttemplates from which we select a subset for the final recognizer.In experiments we evaluate different sizes and types of components andthree standard techniques for component selection. The component classifiersare finally compared to global classifiers on a database of fourobjects.

Word frequency and age of acquisition (AoA) influence word and object recognition and naming. High frequency and early acquired items are processed faster than low frequency and/or late acquired items. The high correlation between word frequency and AoA make these effects difficult to distinguish. However, this difficulty can be avoided by investigating the effects of AoA in the domain of recognising and naming famous faces and names. Face processing a suitable domain because the functional models of face processing were developed by analogy to word and object processing models. Nine experiments on the effects of AoA on face and name processing are reported. Experiment 1 investigated the influence of variables on naming famous faces. The variables were regressed on the speed and accuracy of face naming. Only familiarity and AoA significantly predicted successful naming. A factorial analysis and full replication revealed a consistent advantage for name production to early acquired celebrities' faces (Experiments 2 & 3). Furthermore this advantage was apparent from the first presentation (Experiment 4).Faster face and name recognition occured for early acquired than late acquired celebrities (Experiments 5 & 8). Early acquired names were read aloud faster than late acquired names (Experiment 7). Conversly semantic classifications were made faster to late acquired celebrities' faces (Experiment 6), but there was no effect in the same task to written names (Experiment 9).An effect of AoA for celebrities, whose names are acquired later in life than object names is problematic for the developmental account of AoA. Effects of AoA in recognition tasks are problematic for theorists who propose that speech output is the locus of AoA. A mechanism is proposed to account for the empirical findings. The data also presents a challenge for computer modelling to simulate the combined effects of AoA and cumulative frequency.