Pattern Recognition Phase Research Articles

Self-supervised motor imagery EEG recognition model based on 1-D MTCNN-LSTM network

Objective. Aiming for the research on the brain–computer interface (BCI), it is crucial to design a MI-EEG recognition model, possessing a high classification accuracy and strong generalization ability, and not relying on a large number of labeled training samples. Approach. In this paper, we propose a self-supervised MI-EEG recognition method based on self-supervised learning with one-dimensional multi-task convolutional neural networks and long short-term memory (1-D MTCNN-LSTM). The model is divided into two stages: signal transform identification stage and pattern recognition stage. In the signal transform recognition phase, the signal transform dataset is recognized by the upstream 1-D MTCNN-LSTM network model. Subsequently, the backbone network from the signal transform identification phase is transferred to the pattern recognition phase. Then, it is fine-tuned using a trace amount of labeled data to finally obtain the motion recognition model. Main results. The upstream stage of this study achieves more than 95% recognition accuracy for EEG signal transforms, up to 100%. For MI-EEG pattern recognition, the model obtained recognition accuracies of 82.04% and 87.14% with F1 scores of 0.7856 and 0.839 on the datasets of BCIC-IV-2b and BCIC-IV-2a. Significance. The improved accuracy proves the superiority of the proposed method. It is prospected to be a method for accurate classification of MI-EEG in the BCI system.

Exactness of Mean-Field Equations for Open Dicke Models with an Application to Pattern Retrieval Dynamics.

Open quantum Dicke models are paradigmatic systems for the investigation of light-matter interaction in out-of-equilibrium quantum settings. Albeit being structurally simple, these models can show intriguing physics. However, obtaining exact results on their dynamical behavior is challenging, since it requires the solution of a many-body quantum system with several interacting continuous and discrete degrees of freedom. Here, we make a step forward in this direction by proving the validity of the mean-field semiclassical equations for open multimode Dicke models, which, to the best of our knowledge, so far has not been rigorously established. We exploit this result to show that open quantum multimode Dicke models can behave as associative memories, displaying a nonequilibrium phase transition toward a pattern-recognition phase.

Image Processing-Based Pitting Corrosion Detection Using Metaheuristic Optimized Multilevel Image Thresholding and Machine-Learning Approaches

Pitting corrosion can lead to critical failures of infrastructure elements. Therefore, accurate detection of corroded areas is crucial during the phase of structural health monitoring. This study aims at developing a computer vision and data-driven method for automatic detection of pitting corrosion. The proposed method is an integration of the history-based adaptive differential evolution with linear population size reduction (LSHADE), image processing techniques, and the support vector machine (SVM). The implementation of the LSHADE metaheuristic in this research is multifold. This optimization algorithm is employed in the task of multilevel image thresholding to extract regions of interest from the metal surface. Image texture analysis methods of statistical measurements of color channels, gray-level co-occurrence matrix, and local binary pattern are used to compute numerical features subsequently employed by the SVM-based pattern recognition phase. In addition, the LSHADE metaheuristic is also used to optimize the hyperparameters of the machine-learning approach. Experimental results supported by statistical test points out that the newly developed approach can attain a good predictive result with classification accurate rate = 91.80%, precision = 0.91, recall = 0.94, negative predictive value = 0.93, and F1 score = 0.92. Thus, the newly developed method can be a promising tool to be used in a periodic structural health survey.

A neural network gravitational arc finder based on the Mediatrix filamentation method

Automated arc detection methods are needed to scan the ongoing and next-generation wide-field imaging surveys, which are expected to contain thousands of strong lensing systems. Arc finders are also required for a quantitative comparison between predictions and observations of arc abundance. Several algorithms have been proposed to this end, but machine learning methods have remained as a relatively unexplored step in the arc finding process. In this work we introduce a new arc finder based on pattern recognition, which uses a set of morphological measurements derived from the Mediatrix Filamentation Method as entries to an Artificial Neural Network (ANN). We show a full example of the application of the arc finder, first training and validating the ANN on simulated arcs and then applying the code on four Hubble Space Telescope (HST) images of strong lensing systems. The simulated arcs use simple prescriptions for the lens and the source, while mimicking HST observational conditions. We also consider a sample of objects from HST images with no arcs in the training of the ANN classification. We use the training and validation process to determine a suitable set of ANN configurations, including the combination of inputs from the Mediatrix method, so as to maximize the completeness while keeping the false positives low. In the simulations the method was able to achieve a completeness of about 90% with respect to the arcs that are input to the ANN after a preselection. However, this completeness drops to $\sim$ 70% on the HST images. The false detections are of the order of 3% of the objects detected in these images. The combination of Mediatrix measurements with an ANN is a promising tool for the pattern recognition phase of arc finding. More realistic simulations and a larger set of real systems are needed for a better training and assessment of the efficiency of the method.

Monitoring of mineral processing systems by using textural image analysis

In the last few decades, developments in machine vision technology have led to innovative approaches to the control and monitoring of mineral processing systems. Image representation plays an important role in the performance of the recognition systems used in these approaches, where the use of feature representations based on second-order statistics of the image pixels have predominated. However, these representations may not adequately capture or express the visual textural structure associated with the observed patterns in images. In this study, the use of texton and complex multiscale wavelet representations (steerable pyramids) that exploit higher-order statistical regularities, is investigated. These techniques are applied to two image data sets: industrial platinum group metals froth flotation, and coal particles on a conveyor belt. Compared to grey level co-occurrence matrix and classical wavelet representations, these are observed to improve performance when used as input in the pattern recognition phase.

Automatic recognition of woven fabrics based on texture and using SVM

This paper proposes the recognition and classification of three dominant patterns of woven fabrics such as twill, satin and plain. The proposed classifier is based on the texture analysis of woven fabric images for the recognition. In the pattern recognition phase, three methods are tested and compared: Gabor wavelet, local binary pattern operators and gray-level co-occurrence matrices (GLCM). Taking advantage of the differences between the woven fabric textures, we adopt a technique which is based on the texture of the images in the pattern recognition phase. For the classification phase we used a support vector machine, which we have proven is a suitable classifier for this type of problem. The experimental results show that some of the studied methods are more compatible with this classification problem than others. Although it is the oldest method, GLCM always remains accurate (97.2%). The fusion of the Gabor wavelet and GLCM gives the best result (98%), but GLCM have the better running time.

Novel acoustic features for speech emotion recognition

This paper focuses on acoustic features that effectively improve the recognition of emotion in human speech. The novel features in this paper are based on spectral-based entropy parameters such as fast Fourier transform (FFT) spectral entropy, delta FFT spectral entropy, Mel-frequency filter bank (MFB) spectral entropy, and Delta MFB spectral entropy. Spectral-based entropy features are simple. They reflect frequency characteristic and changing characteristic in frequency of speech. We implement an emotion rejection module using the probability distribution of recognized-scores and rejected-scores. This reduces the false recognition rate to improve overall performance. Recognized-scores and rejected-scores refer to probabilities of recognized and rejected emotion recognition results, respectively. These scores are first obtained from a pattern recognition procedure. The pattern recognition phase uses the Gaussian mixture model (GMM). We classify the four emotional states as anger, sadness, happiness and neutrality. The proposed method is evaluated using 45 sentences in each emotion for 30 subjects, 15 males and 15 females. Experimental results show that the proposed method is superior to the existing emotion recognition methods based on GMM using energy, Zero Crossing Rate (ZCR), linear prediction coefficient (LPC), and pitch parameters. We demonstrate the effectiveness of the proposed approach. One of the proposed features, combined MFB and delta MFB spectral entropy improves performance approximately 10% compared to the existing feature parameters for speech emotion recognition methods. We demonstrate a 4% performance improvement in the applied emotion rejection with low confidence score.

Automatic Classification of Woven Fabrics using Multi-class Support Vector Machine

This paper proposes the recognition and classification of three mean woven fabrics, twill, satin and plain. The proposed classifier is based on the texture analysis of woven fabric images for the recognition. In the pattern recognition phase, three methods are tested and compared: Gabor wavelet, Local Binary Pattern operators (LBP) and gray-level co-occurrence matrices (GLCM). Taking advantage of the difference between the woven fabric textures, we adopt a technique which is based on the texture of the images in the pattern recognition phase. For the classification phase we used a support vector machine (SVM) which we have proven is a suitable classifier for this type of problem The experimental results show that some of the studied methods are more compatible with this classification problem than others. Although it is the oldest method, GLCM always remains accurate (97.2 %).The fusion of the Gabor wavelet and GLCM give the best result (98%), but the GLCM have the better running time.

Adaptive Electrocardiogram Feature Extraction on Distributed Embedded Systems

Tiny embedded systems have not been an ideal outfit for high performance computing due to their constrained resources-limitations in processing power, battery life, communication bandwidth, and memory constrain the applicability of existing complex medical analysis algorithms such as the electrocardiogram (ECG) analysis. Among various limitations, battery lifetime has been a major key technological constraint. In this paper, we address the issue of partitioning such a complex algorithm while the energy consumption due to wireless transmission is minimized. ECG analysis algorithms normally consist of preprocessing, pattern recognition, and classification. Considering the orientation of the ECG leads, we devise a technique to perform preprocessing and pattern recognition locally in small embedded systems attached to the leads. The features detected in the pattern recognition phase are considered for the classification. Ideally, if the features detected for each heartbeat reside in a single processing node, the transmission will be unnecessary. Otherwise, to perform classification, the features must be gathered on a local node and, thus, the communication is inevitable. We perform such a feature grouping by modeling the problem as a hypergraph and applying partitioning schemes which yield a significant power saving in wireless communications. Furthermore, we utilize dynamic reconfiguration by software module migration. This technique, with respect to partitioning, enhances the overall power saving in such systems. Moreover, it adaptively alters the system configuration in various environments and on different patients. We evaluate the effectiveness of our proposed techniques on MIT/BIH benchmarks and, on average, achieve 70 percent energy saving

A study of the robustness of pseudorandom binary-array-based surface characterization

In this paper me report on aspects of the use of color-coded pseudorandom binary arrays (PRBA's) to model three-dimensional scenes. We evaluate the accuracy and the robustness of the pattern recognition phase in the process. Emphasis is on the added value of the use of PRBA's as a tool to make the image processing robust and highly noise insensitive.

A neural network approach for damage detection and identification of structures

This study examines the feasibility of using artificial neural network in conjunction with system identification techniques to detect the existence and to identify the characteristics of damage in composite structures. The methodology proposed here includes a training phase and a recognition phase. In the training phase, candidate models for structures with various types of damage are designated as the patterns. These patterns are organized into pattern classes according to the location and the severity of the damage. Then system identifications are performed to extract the transfer functions as the features of the structural systems. These transfer functions are fed into a multi-layer perceptron (MLP) as the input patterns for training. The MLP serves as a nearest neighborhood classifier. In the pattern recognition phase, a structure with unforeseen damage is classified within the closest class in the training set and the damage in the structure is identified as that of the class. The results of numerical tests demonstrate the feasibility of the proposed method.

A neurobiological model of visual attention and invariant pattern recognition based on dynamic routing of information

We present a biologically plausible model of an attentional mechanism for forming position- and scale-invariant representations of objects in the visual world. The model relies on a set of control neurons to dynamically modify the synaptic strengths of intracortical connections so that information from a windowed region of primary visual cortex (V1) is selectively routed to higher cortical areas. Local spatial relationships (i.e., topography) within the attentional window are preserved as information is routed through the cortex. This enables attended objects to be represented in higher cortical areas within an object-centered reference frame that is position and scale invariant. We hypothesize that the pulvinar may provide the control signals for routing information through the cortex. The dynamics of the control neurons are governed by simple differential equations that could be realized by neurobiologically plausible circuits. In preattentive mode, the control neurons receive their input from a low-level "saliency map" representing potentially interesting regions of a scene. During the pattern recognition phase, control neurons are driven by the interaction between top-down (memory) and bottom-up (retinal input) sources. The model respects key neurophysiological, neuroanatomical, and psychophysical data relating to attention, and it makes a variety of experimentally testable predictions.

Mathematical models for mineral exploration

This papers shows a method whereby different factors and aspects of exploration can be transformed into mathematical models. Selected examples illustrate the procedures followed. Decisions on the basis of uncertain objectives have to be made long before the observations in the field can be started. In a pattern recognition phase, criteria are extracted and characterized on the basis of more or less representative, random samples collected in the surroundings of known mineral deposits. It is not certain the geological model of this training area can be transposed to the prospective areas. In a preparation and planning phase, it has to be proved whether or not a balanced and controlled risk can be expected from a fixed sampling design, the choice of criteria, and the accuracy of measurements performed. By applying mathematical techniques the quality of the decision making can be optimized. In special studies, the probability of design in the vicinity of the undiscovered deposit can be estimated from the observations and the quality of estimation can be controlled in an accurate way.