Catálogo de publicaciones - libros

This paper proposes an approach to reduce the stochastic parsing time with stochastic context-free grammars. The basic idea consists of storing a set of precomputed problems. These precomputed problems are obtained off line from a training corpus or they are computed on line from a test corpus. In this work, experiments with the UPenn Treebank are reported in order to show the performance of both alternatives.

This paper describes a computational framework developed for the extraction of low-level directional primitives present in an image, and subsequent organization through a line segment detector. The system is divided in three stages: extraction of the directional features in the image through an efficient implementation of Gabor wavelet decomposition; reduction of these high dimensionality results by means of a growing cell structure; and extraction of the segments from the image. This last step was first implemented through a pseudo-color Fuzzy Hough Transform and then improved through some principles of the Burns segment detector.

We present a survey of the application of ordinary Kriging to texture interpolation using a variety of models that have been proposed to model the variogram of the image. The novelty of our approach is in the fully automated process of fitting the models to the data over a finite range of values.

Grouping and segmentation of images remains a challenging problem in computer vision. Recently, a number of authors have demonstrated a good performance on this task using spectral methods that are based on the eigensolution of a similarity matrix. In this paper, we implement a variation of the existing methods that combines aspects from several of the best-known eigenvector segmentation algorithms to produce a discrete optimal solution of the relaxed continuous eigensolution.

We present in this paper the application of mathematical morphology operators through a transformation of the Cartesian image into another geometric space, i.e. pol-log image. The conversion into polar-logarithmic coordinates as well as the derived cyclic morphology provides satisfying results in image analysis applied to round objects or spheroid-shaped 3D-object models. As an example of application, an algorithm for the shape analysis of the shape of red blood cells is given.

Hyperspectral applications in remote sensing are often focused on determining the so-called spectral signatures, i.e., the reflectances of materials present in the scene (endmembers) and the corresponding abundance fractions at each pixel in a spatial area of interest. The determination of the number of endmembers in a scene without any prior knowledge is crucial to the success of hyperspectral image analysis. This paper proposes a new mean squared error approach to determine the signal subspace in hyperspectral imagery. The method first estimates the signal and noise correlations matrices, then it selects the subset of eigenvalues that best represents the signal subspace in the least square sense.

Texture-based pixel classification has been traditionally carried out by applying texture feature extraction methods that belong to a same family (e.g., Gabor filters). However, recent work has shown that such classification tasks can be significantly improved if multiple texture methods from different families are properly integrated. In this line, this paper proposes a new selection scheme that automatically determines a subset of those methods whose integration produces classification results similar to those obtained by integrating all the available methods but at a lower computational cost. Experiments with real complex images show that the proposed selection scheme achieves better results than well-known feature selection algorithms, and that the final classifier outperforms recognized texture classifiers.

The processing, description and recognition of dynamic (time-varying) textures are new exciting areas of texture analysis. Many real-world textures are dynamic textures whose retrieval from a video database should be based on both dynamic and static features. In this article, a method for extracting features revealing fundamental properties of dynamic textures is presented. These features are based on the normal flow and on the texture regularity though the sequence. Their discriminative ability is then successfully demonstrated on a full classification process.

Accurately and automatically detecting image orientation is a task of great importance in intelligent image processing. In this paper, we present automatic image orientation detection algorithms based on these features: color moments; harris corner; phase symmetry; edge direction histogram. The statistical learning support vector machines, AdaBoost, Subspace classifier are used in our approach as classifiers. We use Borda Count as combination rule for these classifiers. Large amounts of experiments have been conducted, on a database of more than 6,000 images of real photos, to validate our approaches. Discussions and future directions for this work are also addressed at the end of the paper.

In this paper, a new method for the segmentation of color images is presented. This method searches for an acceptable segmentation of 1D-histograms, according to a “monotone” hypothesis. The algorithm uses recurrence to localize all the modes in the histogram. The algorithm is applied on the hue, saturation and intensity histograms of the image. As a result, an optimal and accurately segmented image is obtained. In contrast to previous state of the art methods uses exclusively the image color histogram to perform segmentation and no spatial information at all.