Pilar Gomez-Gil

FPGA-Based Broken Bars Detection on Induction Motors Under Different Load Using Motor Current Signature Analysis and Mathematical Morphology

Broken bars detection on induction motors has been a topic of interest in recent years. Its detection is important due to the fact that the failure is silent and the consequences it produces as power consumption increasing, vibration, introduction of spurious frequencies in the electric line, among others, can be catastrophic. In this paper, the use of motor current signature analysis and mathematical morphology to detect broken bars on induction motors under different mechanical load condition is analyzed. The proposed algorithm first identifies the motor load and then the motor condition. The statistical analysis of several tests under different motor loads (100%, 75%, 50%, and 25%) and motor condition (healthy, one broken bar, and two broken bars) is presented. The proposed method has been implemented in a field programmable gate array, to be used in real-time online applications. The algorithm obtained in average a 95% accuracy of failure detection.

Acute leukemia classification by ensemble particle swarm model selection

OBJECTIVE Acute leukemia is a malignant disease that affects a large proportion of the world population. Different types and subtypes of acute leukemia require different treatments. In order to assign the correct treatment, a physician must identify the leukemia type or subtype. Advanced and precise methods are available for identifying leukemia types, but they are very expensive and not available in most hospitals in developing countries. Thus, alternative methods have been proposed. An option explored in this paper is based on the morphological properties of bone marrow images, where features are extracted from medical images and standard machine learning techniques are used to build leukemia type classifiers. METHODS AND MATERIALS This paper studies the use of ensemble particle swarm model selection (EPSMS), which is an automated tool for the selection of classification models, in the context of acute leukemia classification. EPSMS is the application of particle swarm optimization to the exploration of the search space of ensembles that can be formed by heterogeneous classification models in a machine learning toolbox. EPSMS does not require prior domain knowledge and it is able to select highly accurate classification models without user intervention. Furthermore, specific models can be used for different classification tasks. RESULTS We report experimental results for acute leukemia classification with real data and show that EPSMS outperformed the best results obtained using manually designed classifiers with the same data. The highest performance using EPSMS was of 97.68% for two-type classification problems and of 94.21% for more than two types problems. To the best of our knowledge, these are the best results reported for this data set. Compared with previous studies, these improvements were consistent among different type/subtype classification tasks, different features extracted from images, and different feature extraction regions. The performance improvements were statistically significant. We improved previous results by an average of 6% and there are improvements of more than 20% with some settings. In addition to the performance improvements, we demonstrated that no manual effort was required during acute leukemia type/subtype classification. CONCLUSIONS Morphological classification of acute leukemia using EPSMS provides an alternative to expensive diagnostic methods in developing countries. EPSMS is a highly effective method for the automated construction of ensemble classifiers for acute leukemia classification, which requires no significant user intervention. EPSMS could also be used to address other medical classification tasks.

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.

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.

Type-2 fuzzy sets applied to pattern matching for the classification of cries of infants under neurological risk

Crying is an acoustic event that contains information about the functioning of the central nervous system, and the analysis of the infants crying can be a support in the distinguishing diagnosis in cases like asphyxia and hyperbilirrubinemia. The classification of baby cry has been intended by the use of different types of neural networks and other recognition approaches. In this work we present a pattern classification algorithm based on fuzzy logic Type 2 with which the classification of infant cry is realized. Experiments as well as results are also shown.

Genetic fuzzy relational neural network for infant cry classification

In this paper we describe a genetic fuzzy relational neural network (FRNN) designed for classification tasks. The genetic part of the proposed system determines the best configuration for the fuzzy relational neural network. Besides optimizing the parameters for the FRNN, the fuzzy membership functions are adjusted to fit the problem. The system is tested in several infant cry database reaching results up to 97.55%. The design and implementation process as well as some experiments along with their results are shown.

A Feature Extraction Method Based on Morphological Operators for Automatic Classification of Leukocytes

In this paper we present preliminary results obtained from the application of morphological operator pecstrum, for the extraction of discriminating characteristics in leukocytes and similar artificial images. Experts have identified six categories of leukocytes, very similar in shape and size, which makes them extremely difficult to distinguish automatically or even by non-expert humans. A feature vector based on a 7-component pecstrum, normalized area, and nucleus-cytoplasm area ratio, was tested using 4 kinds of recognizers: Euclidean distance, k-nearest neighbor, back propagation neural net and support vector machine. Using 36 patterns for training and 18 for testing, recognition of 87% was obtained in the best case, which is encouraging, given the complexity of the problem. The amount of samples used at this point for experiments is not statistically representative, however these results are promising and more experiments will be carried out.

Time series forecasting using recurrent neural networks and wavelet reconstructed signals

In this paper a novel neural network architecture for medium-term time series forecasting is presented. The proposed model, inspired on the Hybrid Complex Neural Network (HCNN) model, takes advantage of information obtained by wavelet decomposition and of the oscillatory abilities of recurrent neural networks (RNN). The prediction accuracy of the proposed architecture is evaluated using 11 economic time series of the NN5 Forecasting Competition for Artificial Neural Networks and Computational Intelligence, obtaining an average SMAPE of 27%. The proposed model shows a better mean performance in time series prediction of 56 values than a feed-forward network and a fully recurrent neural network with a similar number of nodes.

Shape-based hand recognition approach using the morphological pattern spectrum

We propose the use of the morphological pattern spec- trum, or pecstrum, as the base of a biometric shape-based hand recognition system. The system receives an image of the right hand of a subject in an unconstrained pose, which is captured with a commercial flatbed scanner. According to pecstrum property of in- variance to translation and rotation, the system does not require the use of pegs for a fixed hand position, which simplifies the image acquisition process. This novel feature-extraction method is tested using a Euclidean distance classifier for identification and verifica- tion cases, obtaining 97% correct identification, and an equal error rate (EER) of 0.0285 (2.85%) for the verification mode. The obtained results indicate that the pattern spectrum represents a good feature- extraction alternative for low- and medium-level hand-shape-based biometric applications.