Using an ensemble of classifiers, instead of a single classifier, can lead to improved generalization. The gains obtained by combining however, are often affected more by the selection of what is presented to the combiner, than by the actual combining method that is chosen. In this paper we focus on data selection and classifier training methods, in order to "prepare" classifiers for combining. We review a combining framework for classification problems that quantifies the need for reducing the correlation among individual classifiers. Then, we discuss several methods that make the classifiers in an ensemble more complementary. Experimental results are provided to illustrate the benefits and pitfalls of reducing the correlation among classifiers, especially when the training data is in limited supply. 2 1 Introduction A classifier's ability to meaningfully respond to novel patterns, or generalize, is perhaps its most important property (Levin et al., 1990; Wolpert, 1990). In...

Faces represent complex, multidimensional, meaningful visual stimuli and developing a computational model for face recognition is difficult [43]. We present a hybrid neural network solution which compares favorably with other methods. The system combines local image sampling, a self-organizing map neural network, and a convolutional neural network. The self-organizing map provides a quantization of the image samples into a topological space where inputs that are nearby in the original space are also nearby in the output space, thereby providing dimensionality reduction and invariance to minor changes in the image sample, and the convolutional neural network provides for partial invariance to translation, rotation, scale, and deformation. The convolutional network extracts successively larger features in a hierarchical set of layers. We present results using the Karhunen-Loeve transform in place of the self-organizing map, and a multi-layer perceptron in place of the convolutional netwo...

We consider the general problem of utilizing both labeled and unlabeled data to improve classification accuracy. Under the assumption that the data lie on a submanifold in a high dimensional space, we develop an algorithmic framework to classify a partially labeled data set in a principled manner. The central idea of our approach is that classification functions are naturally defined only on the submanifold in question rather than the total ambient space. Using the Laplace-Beltrami operator one produces a basis (the Laplacian Eigenmaps) for a Hilbert space of square integrable functions on the submanifold. To recover such a basis, only unlabeled examples are required. Once such a basis is obtained, training can be performed using the labeled data set. Our algorithm models the manifold using the adjacency graph for the data and approximates the Laplace-Beltrami operator by the graph Laplacian. We provide details of the algorithm, its theoretical justification, and several practical applications for image, speech, and text classification. 1.

A nonlinear black box structure for a dynamical system is a model structure that is prepared to describe virtually any nonlinear dynamics. There has been considerable recent interest in this area with structures based on neural networks, radial basis networks, wavelet networks, hinging hyperplanes, as well as wavelet transform based methods and models based on fuzzy sets and fuzzy rules. This paper describes all these approaches in a common framework, from a user's perspective. It focuses on what are the common features in the different approaches, the choices that have to be made and what considerations are relevant for a successful system identification application of these techniques. It is pointed out that the nonlinear structures can be seen as a concatenation of a mapping from observed data to a regression vector and a nonlinear mapping from the regressor space to the output space. These mappings are discussed separately. The latter mapping is usually formed as a basis function e...

Combining learning with vision techniques in interactive image retrieval has been an active research topic during the past few years. However, existing learning techniques either are based on heuristics or fail to analyze the working conditions. Furthermore, there is almost no in depth study on how to effectively learn from the users when there are multiple visual features in the retrieval system. To address these limitations, in this paper, we present a vigorous optimization formulation of the learning process and solve the problem in a principled way. By using Lagrange multipliers, we have derived explicit solutions, which are both optimal and fast to compute. Extensive comparisons against state-ofthe-art techniques have been performed. Experiments were carried out on a large-size heterogeneous image collection consisting of 17,000 images. Retrieval performance was tested under a wide range of conditions. Various evaluation criteria, including precision-recall curve and rank measure, have demonstrated the effectiveness and robustness of the proposed technique. 1.

The development of an EEG-based brain-computer interface (BCI) requires rapid and reliable discrimination of EEG patterns, e. g., associated with motor imagery. One sided hand movement imagination results in EEG changes located at contra- and ipsilateral central areas. We demonstrate that spatial filters for multi-channel EEG effectively extract discriminatory information from two populations of single-trial EEG, recorded during left and right hand movement imagery. The best classification results for three subjects are 90.8%, 92.7%, and 99.7%. The spatial filters are estimated from a set of data by the method of Common Spatial Patterns and reflect the specific activation of cortical areas. The method performs a weighting of the electrodes according to their importance for the classification task. The high recognition rates and computational simplicity make it a promising method for an EEG-based brain-computer interface.

Finding structures in vast multidimensional data sets, be they measurement data, statistics, or textual documents, is difficult and time-consuming. Interesting, novel relations between the data items may be hidden in the data. The selforganizing map (SOM) algorithm of Kohonen can be used to aid the exploration: the structures in the data sets can be illustrated on special map displays. In this work, the methodology of using SOMs for exploratory data analysis or data mining is reviewed and developed further. The properties of the maps are compared with the properties of related methods intended for visualizing highdimensional multivariate data sets. In a set of case studies the SOM algorithm is applied to analyzing electroencephalograms, to illustrating structures of the standard of living in the world, and to organizing full-text document collections. Measures are proposed for evaluating the quality of different types of maps in representing a given data set, and for measuring the robu...

Abstract not found

We present a novel approach for recognizing faces in images taken from different directions and under different illumination. The method is based on a 3D morphable face model that encodes shape and texture in terms of model parameters, and an algorithm that recovers these parameters from a single image of a face. For face identification, we use the shape and texture parameters of the model that are separated from imaging parameters, such as pose and illumination. In addition to the identity, the system provides a measure of confidence. We report experimental results for more than 4000 images from the publicly available CMU-PIE database. 1

Abstract not found