Catálogo de publicaciones - libros

The wave is a complex and important phenomenon for structures de signs in the coastal zones and beaches. This paper presents a novel system for the generation of spectral patterns of unidirectional irregular waves in research laboratories. The system’s control basic elements are a linear motor, a servo controller and a personal computer. The used main mathematical tools are a feed forward neural network, digital signal processing and statistical analysis. The research aim is to obtain a system of more accuracy and small response time. This behavior is interpreted, in marine hydraulics, as a fast calibration of experiments. The wave power spectrums are generated in a test channel of rectangular section with dimensions: length 12 m; depth 40 cm; width 30 cm.

A simple note onset detection system for music is presented in this work. To detect onsets, a 1/12 octave filterbank is simulated in the frequency domain and the band derivatives in time are considered. The first harmonics of a tuned instrument are close to the center frequency of these bands and, in most instruments, these harmonics are those with the highest amplitudes. The goal of this work is to make a musically motivated system which is sensitive on onsets in music but robust against the spectrum variations that occur at times that do not represent onsets. Therefore, the system tries to find semitone variations, which correspond to note onsets. Promising results are presented for this real time onset detection system.

In this work we propose to monitor the cutting tool-wear condition in a CNC-machining center by using continuous Hidden Markov Models (HMM). A database was built with the vibration signals obtained during the machining process. The workpiece used in the milling process was aluminum 6061. Cutting tests were performed on a Huron milling machine equipped with a Sinumerik 840D open CNC. We trained/tested the HMM under 18 different operating conditions. We identified three key transitions in the signals. First, the cutting tool touches the workpiece. Second, a stable waveform is observed when the tool is in contact with the workpiece. Third, the tool finishes the milling process. Considering these transitions, we use a five-state HMM for modeling the process. The HMMs are created by preprocessing the waveforms, followed by training step using Baum-Welch algorithm. In the recognition process, the signal waveform is also preprocessed, then the trained HMM are used for decoding. Early experimental results validate our proposal in exploiting speech recognition frameworks in monitoring machining centers. The classifier was capable of detecting the cutting tool condition within large variations of spindle speed and feed rate, and accuracy of 84.19%.

A new method for hand gesture recognition is proposed which is based on an innovative Self-Growing and Self-Organized Neural Gas (SGONG) network. Initially, the region of the hand is detected by using a color segmentation technique that depends on a skin-color distribution map. Then, the SGONG network is applied on the segmented hand so as to approach its topology. Based on the output grid of neurons, palm geometric characteristics are obtained which in accordance with powerful finger features allow the identification of the raised fingers. Finally, the hand gesture recognition is accomplished through a probability-based classification method.

As an efficient sparse coding and feature extraction method, independent component analysis (ICA) has been widely used in speech signal processing. In this paper, ICA method is studied in extracting low frequency features of underwater acoustic signals. The generalized Gaussian model (GGM) is introduced as the p.d.f. estimator in ICA to extract the basis vectors. It is demonstrated that the ICA features of ship radiated signals are localized both in time and frequency domain. Based on the ICA features, an extended de-noising method is proposed for underwater acoustic signals which can extract the basis vectors directly from the noisy observation. The de-noising experiments of underwater acoustic signals show that the proposed method offers an efficient approach for detecting weak underwater acoustic signals from noise environment.

In this paper we study the ICA feature extraction method for Chinese speech signals. The generalized Gaussian model (GGM) is introduced as the p.d.f. estimator in ICA since it can provide a general method for modeling non-Gaussian statistical structure of univariate distributions. It is demonstrated that the ICA features of Chinese speech are localized in both time and frequency domain and the resulting coefficients are statistically independent and sparse. The GGM-based ICA method is also used in extracting the basis vectors directly from the noisy observation, which is an efficient method for noise reduction when priori knowledge of source data is not acquirable. The de-nosing experiments show that the proposed method is more efficient than conventional methods in the environment of additive white Gaussian noise.

Wavelets are widely used in numerous applied fields involving for example signal analysis, image compression or function approximation. The idea of adapting wavelet to specific problems, it means to create and use problem and data dependent wavelets, has been developed for various purposes. In this paper, we are interested in to define, starting from a given pattern, an efficient design of FIR adapted wavelets based on the lifting scheme. We apply the constructed wavelet for pattern detection in the 1D case. To do so, we propose a three stages detection procedure which is finally illustrated by spike detection in EEG.

In this paper we present a method to detect edges in images. The method consists of using a 3x3 pixel mask to scan the image, moving it from left to right and from top to bottom, one pixel at a time. Each time it is placed on the image, an agglomerative hierarchical cluster analysis is applied to the eight outer pixels. When there is more than one cluster, it means that window is on an edge, and the central pixel is marked as an edge point. After scanning all the image, we obtain a new image showing the marked pixels around the existing edges of the image. Then a thinning algorithm is applied so that the edges are well defined. The method results to be particularly efficient when the image is contaminated. In those cases, a previous restoration method is applied.

In this paper a new method for edge detection in grayscale images is presented. It is based on the use of the Kohonen self-organizing map (SOM) neural network combined with the methodology of Canny edge detector. Gradient information obtained from different masks and at different smoothing scales is classified in three classes (Edge, Non Edge and Fuzzy Edge) using an hierarchical Kohonen network. Using the three classes obtained, the final stage of hysterisis thresholding is performed in a fully automatic way. The proposed technique is extensively tested with success.

The proposed technique deals with jigsaw puzzles and takes advantage of both geometrical and color features. It is considered that an image is being divided into pieces. The shape of these pieces is not predefined, yet the background’s color is. The whole method concerns a recurrent algorithm, which initially, finds the most important corner points around the contour of a piece, afterwards performs color segmentation with a Kohonen’s SOFM based technique and finally uses a comparing routine. This routine is based on the corner points found before. It compares a set of angles, the color of the image around the region of the corner points, the color of the contour and finally compares sequences of points by calculating the Euclidean distance of luminance between them. At a final stage the method decides which pieces match. If the result is not satisfying, the algorithm is being repeated with new adaptive modified parameter values as far as the corner points and the color segmentation is concerned.