Vicente Alarcon-Aquino

Multiresolution FIR neural-network-based learning algorithm applied to network traffic prediction

In this paper, a multiresolution finite-impulse-response (FIR) neural-network-based learning algorithm using the maximal overlap discrete wavelet transform (MODWT) is proposed. The multiresolution learning algorithm employs the analysis framework of wavelet theory, which decomposes a signal into wavelet coefficients and scaling coefficients. The translation-invariant property of the MODWT allows alignment of events in a multiresolution analysis with respect to the original time series and, therefore, preserving the integrity of some transient events. A learning algorithm is also derived for adapting the gain of the activation functions at each level of resolution. The proposed multiresolution FIR neural-network-based learning algorithm is applied to network traffic prediction (real-world aggregate Ethernet traffic data) with comparable results. These results indicate that the generalization ability of the FIR neural network is improved by the proposed multiresolution learning algorithm.

A comparative simulation study of wavelet based denoising algorithms

In this paper, we present a comparative simulation study of three denoising algorithms using wavelets. The denoising algorithms (i.e., universal threshold, minimax threshold and rigorous SURE threshold) have been used to remove white Gaussian noise from synthetic and real signals. The analysis is done by applying soft and hard thresholds to signals with different sample sizes. The mean squared error (MSE) is used to evaluate the performance of these algorithms. The results show that the rigorous SURE algorithm with a hard threshold has a better performance than other algorithms in synthetic signals. On the other hand, the universal threshold algorithm with a soft threshold shows the best performance in real signals when using the Daubechies wavelet with 5 vanishing moments.

On Signal P-300 Detection for BCI Applications Based on Wavelet Analysis and ICA Preprocessing

This paper describes an experiment on the detection of a P-300 rhythm from electroencephalographic signals for brain computer interfaces applications. The P300 evoked potential is obtained from visual stimuli followed by a motor response from the subject. The EEG signals are obtained with a 14 electrodes Emotiv EPOC headset. Preprocessing of the signals includes denoising and blind source separation using an Independent Component Analysis algorithm. The P300 rhythm is detected through a time-scale analysis based on the discrete wavelet transform (DWT). Comparison using the Short Time Fourier Transform (STFT), and Wigner–Ville Distribution (WVD) indicates that the DWT outperforms the others as an analyzing tool for P300 rhythm detection.

Single-Step Prediction of Chaotic Time Series Using Wavelet-Networks

This paper presents a wavelet neural-network for chaotic time series prediction. Wavelet-networks are inspired by both the feed-forward neural network and the theory underlying wavelet decompositions. Wavelet-networks are a class of neural network that take advantage of good localization properties of multiresolution analysis and combine them with the approximation abilities of neural networks. This kind of networks uses wavelets as activation functions in the hidden layer and a type of backpropagation algorithm is used for its learning. Comparisons are made between a wavelet-network and the typical feedforward network trained with the back-propagation algorithm. The results reported in this paper show that wavelet-networks have better prediction properties than its similar back-propagation networks

Breaking text-based CAPTCHAs with variable word and character orientation

A novel approach for automatic segmentation and recognition of CAPTCHAs with variable orientation and random collapse of overlapped characters is presented in this paper. Additionally, the extension of the proposed approach to break reCAPTCHA of version of 2012 is also discussed. The original proposal consists in straightening characters and word in CAPTCHA exploiting then a three-color bar code for their segmentation. The recognition of straightened characters and whole word is provided by the proposed original SVM-based learning classifier. The main goal of this research is to reduce vulnerability of CAPTCHA from spam and frauds as well as to provide an approach for recognizing either handwritten or degraded and damaged texts in ancient manuscripts by OCR systems. The designed framework for breaking CAPTCHAs by the proposed approach has been tested achieving average segmentation success rate up to 82% for reCAPTCHA of version 2011 and achieving 95.5% by extended approach for reCAPTCHA of version 2012 with response time less than 0.5s per two-word reCAPTCHA. The implemented SVM classifier shows a competitive precision about 94%. The obtained very satisfactory results confirm that the proposed approach may be used for development of new security mechanisms to protect users against cyber-criminal activities and Internet threats. Automatic segmentation and recognition of CAPTCHAs in Web sites is proposed.Anti-recognition techniques use collapsed characters with variable orientation.Aligned word and straightened characters are segmented by three-color bar code.Original SVM-based learning classifier provides real-time CAPTCHA recognition.Extended approach for beating reCAPTCHA of version 2012 shows better performance.

Wavelet-based smoke detection in outdoor video sequences

In this paper an approach to detect smoke columns from outdoor forest video sequences is proposed. The approach follows three basic steps. The first step is an image pre-processing block which resizes the image by applying a bicubic interpolation algorithm. The image is then transformed to its intensity values with a gray-scale transformation and finally the image is grouped by common areas with an image indexation. The second step consists of a smoke detection algorithm which performs a stationary wavelet transform (SWT) to remove high frequencies on horizontal, vertical, and diagonal details. The inverse SWT is then implemented and finally the image is compared to a non-smoke scene in order to determine the possible regions of interest (ROI). In order to reduce the number of false alarms, the final step of the proposed approach consists on a smoke verification algorithm, which determines whether the ROI is increasing its area or not. These results are combined to reach a final decision for detecting a smoke column on a sequence of static images from an outdoor video. Experimental results show that multi-resolution wavelet analysis is more accurate than the traditional low-pass filters on this application.

A Neural Network Scheme for Long-Term Forecasting of Chaotic Time Series

The accuracy of a model to forecast a time series diminishes as the prediction horizon increases, in particular when the prediction is carried out recursively. Such decay is faster when the model is built using data generated by highly dynamic or chaotic systems. This paper presents a topology and training scheme for a novel artificial neural network, named “Hybrid-connected Complex Neural Network” (HCNN), which is able to capture the dynamics embedded in chaotic time series and to predict long horizons of such series. HCNN is composed of small recurrent neural networks, inserted in a structure made of feed-forward and recurrent connections and trained in several stages using the algorithm back-propagation through time (BPTT). In experiments using a Mackey-Glass time series and an electrocardiogram (ECG) as training signals, HCNN was able to output stable chaotic signals, oscillating for periods as long as four times the size of the training signals. The largest local Lyapunov Exponent (LE) of predicted signals was positive (an evidence of chaos), and similar to the LE calculated over the training signals. The magnitudes of peaks in the ECG signal were not accurately predicted, but the predicted signal was similar to the ECG in the rest of its structure.

A motor imagery BCI experiment using wavelet analysis and spatial patterns feature extraction

A brain computer interface (BCI) is a system that aims to control devices by analyzing brain signals patterns. In this work, a convenient time-frequency representation (TFR) for visualizing ERD/ERS phenomenon (Event related synchronization and desynchronization) based on Hilbert transform and spatial patterns is addressed, and a wavelet based feature extraction method for motor imagery tasks is presented. The feature vectors are constructed with four statistical and energy parameters obtained from wavelet decomposition, based on the sub-band coding algorithm. Experimentation with three classification methods for comparison purposes was carried out using Linear Discriminant Analysis (LDA), Quadratic Discriminant Analysis (QDA), and Support Vector Machine (SVM). In each case, ten-fold validation is used to obtain average misclassification rates.

Breaking reCAPTCHAs with unpredictable collapse: heuristic character segmentation and recognition

In this paper we present a novel approach for automatic segmentation and recognition of reCAPTCHA in Web sites. It is based on CAPTCHA image preprocessing with character alignment, morphological segmentation with three-color bar character encoding and heuristic recognition. The original proposal consists in exploiting three-color bar code for characters in CAPTCHA for their robust segmentation with presence of random collapse overlapping letters and distortions by particular patterns of waving rotation. Additionally, a novel implementation of SVM-based learning classifier for recognition of combinations of characters in training corpus has been proposed that permits to increment more than twice the recognition success rate without time extension of system response. The main goal of this research is to reduce vulnerability of CAPTCHA from spam and frauds as well as to provide a novel approach for recognizing either handwritten or degraded and damaged texts in ancient manuscripts. Our designed framework implementing the proposed approach has been tested in real-time applications with sites used CAPTCHAS achieving segmentation success rate about of 82% and recognition success rate about of 94%.

Instance Selection and Feature Weighting Using Evolutionary Algorithms

Machine learning algorithms are commonly used in real-world applications for solving complex problems where it is difficult to get a mathematical model. The goal of machine learning algorithms is to learn an objective function from a set of training examples where each example is defined by a feature set. Regularly, real world applications have many examples with many features; however, the objective function depends on few of them. The presence of noisy examples or irrelevant features in a dataset degrades the performance of machine learning algorithms; such is the case of k-nearest neighbor machine learning algorithm (k-NN). Thus choosing good instance and feature subsets may improve the algorithms performance. Evolutionary algorithms proved to be good techniques for finding solutions in a large solution space and to be stable in the presence of noise. In this work, we address the problem of instance selection and feature weighting for instance-based methods by means of a genetic algorithm (GA) and evolution strategies (ES). We show that combining GA and ES with a k-NN algorithm can improve the predictive accuracy of the resulting classifier