Green mine construction is the main melody of mining development and problems such as safe production, energy saving and consumption reduction need to be solved urgently. The working conditions of the mill are complex in the process of grinding. Aiming at the problems existing in the feature extraction and load prediction of the mill, a signal-processing method based on adaptive chirp mode decomposition (ACMD) and a standardized variable distance classifier (SVD) is proposed. Firstly, the recursive framework of the ACMD method is used to obtain the initial frequency of mill vibration signals. Secondly, the initial frequency is used to reconstruct the high-resolution component of the mill vibration signal through the iterative frame in the ACMD method. The frequency corresponding to the frequency domain peak of the reconstructed signal is then selected as the mill load feature vector. Finally, with consideration to the influence of standard deviation and standardized variable factors on the feature vectors, a standardized variable distance classifier is proposed. The feature vectors of the mill load are input into the SVD model for training, and the state types of the mill load are obtained. The method is applied to the grinding experiment and the results show that the frequency-domain features obtained by the mill vibration signal-processing method based on ACMD-SVD are obvious, which has high accuracy in the identification of mill load types, and provides a new idea for the extraction of mill load features and prediction of the mill load.

The aim of the presented work was the development of a tracking algorithm for a stereoscopic camera setup equipped with an additional inertial sensor. The input of the algorithm consists of the image sequence, angular velocity and linear acceleration vectors measured by the inertial sensor. The main assumption of the project was fusion of data streams from both sources to obtain more accurate ego-motion estimation. An electronic module for recording the inertial sensor data was built. Inertial measurements allowed a coarse estimation of the image motion field that has reduced its search range by standard image-based methods. Continuous tracking of the camera motion has been achieved (including moments of image information loss). Results of the presented study are being implemented in a currently developed obstacle avoidance system for visually impaired pedestrians.

This article shows that Classical Arabic expresses verbal number. Arabic, of all the Semitic language family, meets the typological tests of the languages expressing verbal number. In addition, I will show that Classical Arabic provides a morphological verb form to express number. I will, however, show that for the form to express verbal number it requires a combination of morphological and semantic conditions. Without which the designated form does not express number, but expresses transitivity or the transfer of agency. These conditions are: form II must come from a root that has a form I, form I must be the transitive meaning of the root and the root must express an instant action. Form II, therefore, does not exclusively express number. Verbal number in Arabic is conditional. However, I will also propose that when form II verb expresses number, it does not express the transfer of agency.

In the last decade of the XX-th century, several academic centers have launched intensive research programs on the brain-computer interface (BCI). The current state of research allows to use certain properties of electromagnetic waves (brain activity) produced by brain neurons, measured using electroencephalographic techniques (EEG recording involves reading from electrodes attached to the scalp - the non-invasive method - or with electrodes implanted directly into the cerebral cortex - the invasive method). A BCI system reads the user's “intentions” by decoding certain features of the EEG signal. Those features are then classified and "translated" (on-line) into commands used to control a computer, prosthesis, wheelchair or other device. In this article, the authors try to show that the BCI is a typical example of a measurement and control unit.

In order to achieve accurate identification and segmentation of ore under complex working conditions, machine vision and neural network technology are used to carry out intelligent detection research on ore, an improved Mask RCNN instance segmentation algorithm is proposed. Aiming at the problem of misidentification of stacked ores caused by the loss of deep feature details during the feature extraction process of ore images, an improved Multipath Feature Pyramid Network (MFPN) was proposed. The network firstly adds a single bottom-up feature fusion path, and then adds with the top-down feature fusion path of the original algorithm, which can enrich the deep feature details and strengthen the fusion of the network to the feature layer, and improve the accuracy of the network to the ore recognition. The experimental results show that the algorithm proposed in this paper has a recognition accuracy of 96.5% for ore under complex working conditions, and the recall rate and recall rate function values reach 97.4% and 97.0% respectively, and the AP75 value is 6.84% higher than the original algorithm. The detection results of the ore in the actual scene show that the mask size segmented by the network is close to the actual size of the ore, indicating that the improved network model proposed in this paper has achieved a good performance in the detection of ore under different illumination, pose and background. Therefore, the method proposed in this paper has a good application prospect for stacked ore identification under complex working conditions.

The paper presents two algorithms as a solution to the problem of identifying fraud intentions of a customer. Their purpose is to generate variables that contribute to fraud models’ predictive power improvement. In this article, a novel approach to the feature engineering, based on anomaly detection, is presented. As the choice of statistical model used in the research improves predictive capabilities of a solution to some extent, most of the attention should be paid to the choice of proper predictors. The main finding of the research is that model enrichment with additional predictors leads to the further improvement of predictive power and better interpretability of anti-fraud model. The paper is a contribution to the fraud prediction problem but the method presented may generate variable input to every tool equipped with variableselection algorithm. The cost is the increased complexity of the models obtained. The approach is illustrated on a dataset from one of the European banks.

Efficient, accurate data collection from imagery is the key to an economical generation of useful geospatial products. Incremental developments of traditional geospatial data collection and the arrival of new image data sources cause new software packages to be created and existing ones to be adjusted to enable such data to be processed. In the past, BAE Systems’ digital photogrammetric workstation, SOCET SET ® , met fi n de siècle expectations in data processing and feature extraction. Its successor, SOCET GXP ® , addresses today’s photogrammetric requirements and new data sources. SOCET GXP is an advanced workstation for mapping and photogrammetric tasks, with automated functionality for triangulation, Digital Elevation Model (DEM) extraction, orthorectification and mosaicking, feature extraction and creation of 3-D models with texturing. BAE Systems continues to add sensor models to accommodate new image sources, in response to customer demand. New capabilities added in the latest version of SOCET GXP facilitate modeling, visualization and analysis of 3-D features.

In this paper, a new feature-extraction method is proposed to achieve robustness of speech recognition systems. This method combines the benefits of phase autocorrelation (PAC) with bark wavelet transform. PAC uses the angle to measure correlation instead of the traditional autocorrelation measure, whereas the bark wavelet transform is a special type of wavelet transform that is particularly designed for speech signals. The extracted features from this combined method are called phase autocorrelation bark wavelet transform (PACWT) features. The speech recognition performance of the PACWT features is evaluated and compared to the conventional feature extraction method mel frequency cepstrum coefficients (MFCC) using TI-Digits database under different types of noise and noise levels. This database has been divided into male and female data. The result shows that the word recognition rate using the PACWT features for noisy male data (white noise at 0 dB SNR) is 60%, whereas it is 41.35% for the MFCC features under identical conditions

Keypoint detection is a basic step in many computer vision algorithms aimed at recognition of objects, automatic navigation and analysis of biomedical images. Successful implementation of higher level image analysis tasks, however, is conditioned by reliable detection of characteristic image local regions termed keypoints. A large number of keypoint detection algorithms has been proposed and verified. In this paper we discuss the most important keypoint detection algorithms. The main part of this work is devoted to description of a keypoint detection algorithm we propose that incorporates depth information computed from stereovision cameras or other depth sensing devices. It is shown that filtering out keypoints that are context dependent, e.g. located at boundaries of objects can improve the matching performance of the keypoints which is the basis for object recognition tasks. This improvement is shown quantitatively by comparing the proposed algorithm to the widely accepted SIFT keypoint detector algorithm. Our study is motivated by a development of a system aimed at aiding the visually impaired in space perception and object identification.

Currently the recidivism rate in Ukraine. This indicates failure to achieve the goal of punishment – correction of the convict. The purpose of the article is to research the problems of resocialization of convicts, taking into consideration the psychological characteristics of the person serving the sentence. The subject of research: the subject of research is the resocialization of convicts. The following scientific methods were used to study the international experience of resocialization of convicts, to prove the hypotheses, to formulate conclusions: dialectical method, monographic method, logical method, comparative method, generalization method, system and structural method. The results of the research: it was found out that serving a certain term of imprisonment or life imprisonment affects convicts and leads to a change in their psychology in completely different ways. It is proved that the process of resocialization should be set up during the selection of convict’s type and size of punishment (taking into account the circumstances of the case, the perpetrator personality and criminogenic risks that may contribute to recidivism), continue during punishment (using training, work and communication, and providing psychological support to overcome possible psychological crises) and finish after the release from penitentiary institutions (with control over the released, employment assistance or the provision of temporary residence).

The paper presents qualitative, Bayesian model used 10 determine some interdependencies between sorption features for mineral soils in southern Poland. Sorption properties are very important, crucial for measure or fertility, nutrient retention capacity, and the capacity to protect groundwater from coutaminution. Cation exchange capacity (CFC) is a commonly applied indicator otihc soils conditions or vulncrahilitv. Base saturation (BS) is an important clement of hazard degree assessment in soils lying within reach of impact acidifying agents. The considered soils represented different valuation classes and differed in their typology. The Bayesian model is used lor interdependences assessment.

Complex gaps may be formed when carrying out live working in substations, while the discharge characteristics of complex gaps are different from those of single gaps. This paper focuses on the prediction of critical 50% positive switching impulse breakdown voltage ( U _{50–crit +} of phase-to-phase complex gaps formed in 220 kV substations. Firstly, several electric field features were defined on the shortest discharge path of the complex gap to reflect the electric field distribution. Then support vector machine (SVM) prediction models were established according to the connection between electric field distribution and breakdown voltage. Finally, the U _50–crit¸+ data of the complex gap were obtained through twice electric field calculations and predictions. The prediction results show that the minimum U _{50–crit +} of phase-to-phase complex gaps is 1147 kV, and the critical position is 0.9 m away from the high voltage conductor, accounting for 27% of the whole gap. Both critical position and voltage are in good agreement with the values provided in IEC 61472.

In the field of medicine there is a need for the automatic detection of retinal disorders. Blindness in older persons is primarily caused by Central Retinal Vein Occlusion (CRVO). It results in rapid, irreversible eyesight loss, therefore, it is essential to identify and address CRVO as soon as feasible. Hemorrhages, which can differ in size, pigment, and shape from dot-shaped to flame hemorrhages, are one of the earliest symptoms of CRVO. The early signs of CRVO are, hemorrhages, however, so mild that ophthalmologists must dynamically observe such indicators in the retina image known as the fundus image, which is a challenging and time-consuming task. It is also difficult to segment hemorrhages since the blood vessels and hemorrhages (HE) have the same color properties also there is no particular shape for hemorrhages and it scatters all over the fundus image. A challenging study is needed to extract the characteristics of vein deformability and dilatation. Furthermore, the quality of the captured image affects the efficacy of feature Identification analysis. In this paper, a deep learning approach for CRVO extraction is proposed.

In this paper a review on biometric person identification has been discussed using features from retinal fundus image. Retina recognition is claimed to be the best person identification method among the biometric recognition systems as the retina is practically impossible to forge. It is found to be most stable, reliable and most secure among all other biometric systems. Retina inherits the property of uniqueness and stability. The features used in the recognition process are either blood vessel features or non-blood vessel features. But the vascular pattern is the most prominent feature utilized by most of the researchers for retina based person identification. Processes involved in this authentication system include pre-processing, feature extraction and feature matching. Bifurcation and crossover points are widely used features among the blood vessel features. Non-blood vessel features include luminance, contrast, and corner points etc. This paper summarizes and compares the different retina based authentication system. Researchers have used publicly available databases such as DRIVE, STARE, VARIA, RIDB, ARIA, AFIO, DRIDB, and SiMES for testing their methods. Various quantitative measures such as accuracy, recognition rate, false rejection rate, false acceptance rate, and equal error rate are used to evaluate the performance of different algorithms. DRIVE database provides 100% recognition for most of the methods. Rest of the database the accuracy of recognition is more than 90%.

The paper considers the problem of increasing the generalization ability of classification systems by creating an ensemble of classifiers based on the CNN architecture. Different structures of the ensemble will be considered and compared. Deep learning fulfills an important role in the developed system. The numerical descriptors created in the last locally connected convolution layer of CNN flattened to the form of a vector, are subjected to a few different selection mechanisms. Each of them chooses the independent set of features, selected according to the applied assessment techniques. Their results are combined with three classifiers: softmax, support vector machine, and random forest of the decision tree. All of them do simultaneously the same classification task. Their results are integrated into the final verdict of the ensemble. Different forms of arrangement of the ensemble are considered and tested on the recognition of facial images. Two different databases are used in experiments. One was composed of 68 classes of greyscale images and the second of 276 classes of color images. The results of experiments have shown high improvement of class recognition resulting from the application of the properly designed ensemble.

The paper presents the fusion approach of different feature selection methods in pattern recognition problems. The following methods are examined: nearest component analysis, Fisher discriminant criterion, refiefF method, stepwise fit, Kolmogorov-Smirnov criteria, T2-test, Kruskall-Wallis test, feature correlation with class, and SVM recursive feature elimination. The sensitivity to the noisy data as well as the repeatability of the most important features are studied. Based on this study, the best selection methods are chosen and applied in the process of selection of the most important genes and gene sequences in a dataset of gene expression microarray in prostate and ovarian cancers. The results of their fusion are presented and discussed. The small selected set of such genes can be treated as biomarkers of cancer.

Individual identification of similar communication emitters in the complex electromagnetic environment has great research value and significance in both military and civilian fields. In this paper, a feature extraction method called HVG-NTE is proposed based on the idea of system nonlinearity. The shape of the degree distribution, based on the extraction of HVG degree distribution, is quantified with NTE to improve the anti-noise performance. Then XGBoost is used to build a classifier for communication emitter identification. Our method achieves better recognition performance than the state-of-the-art technology of the transient signal data set of radio stations with the same plant, batch, and model, and is suitable for a small sample size.

The paper discusses a way of choosing the design features (geometry, the rate of grinding and thrust) of ring-ball mills. Various methods of calculating the optimal rate of grinding have been compared. Basing on experimental investigations on the pilot-plant and industrial scale, the influence of the angular velocity and the thrust on the mill have been verified, and the interdependence between the rate of grinding and the thrust of the grinding elements have been explained.

The empirical-analytic model of milling in a ring-ball mill has been presented. It concerns the interaction of the basic design features of the grinding unit (geometry, rate of grinding and thrust of the balls) on the maximum efficiency of the mill. The production of pulverized coal was expressed by the product of the flux of material drawn in by the balls and the so-called "grinding effect of the balls" (defined by the increase of the mass fraction of dust in its flux). The kinematic quantities (among others, the flux of loose material drawn in by the balls) have been calculated on the basis of a simple analytical description of the flow of particles and some parametrical assumptions. The grinding properties of coal have been determined making use of laboratory tests of its cruising by the rollers. Some verifications of the grinding model on the experimental test stand with a ring-ball mill have been presented. The test stand is installed at the Institute of Power Engineering and Turbomachinery of the Silesian University of Technology.

In order to explore the impact of coal and gangue particle size changes on recognition accuracy and to improve the single particle size of coal and gangue identification accuracy of sorting equipment, this study established a database of different particle sizes of coal and gangue through image gray and texture feature extraction, using a relief feature selection algorithm to compare different particle size of coal and gangue optimal features of the combination, and to identify the points and particle size of coal and gangue. The results show that the optimal features and number of coal and gangue are different with different particle sizes. Based on visible-light coal and gangue separation technology, the change of coal and gangue particle size cause fluctuations in the recognition accuracy, and the fluctuation of recognition accuracy will gradually decrease with increases in the number of features. In the process of particle size classification, if the training model has a single particle size range, the recognition accuracy of each particle size range is low, with the highest recognition accuracy being 98% and the average recognition rate being only 97.2%. The method proposed in this paper can effectively improve the recognition accuracy of each particle size range. The maximum recognition accuracy is 100%, the maximum increase is 4%, and the average recognition accuracy is 99.2%. Therefore, this method has a high practical application value for the separation of coal and gangue with single particle size.

Speech emotion recognition is deemed to be a meaningful and intractable issue among a number of do- mains comprising sentiment analysis, computer science, pedagogy, and so on. In this study, we investigate speech emotion recognition based on sparse partial least squares regression (SPLSR) approach in depth. We make use of the sparse partial least squares regression method to implement the feature selection and dimensionality reduction on the whole acquired speech emotion features. By the means of exploiting the SPLSR method, the component parts of those redundant and meaningless speech emotion features are lessened to zero while those serviceable and informative speech emotion features are maintained and selected to the following classification step. A number of tests on Berlin database reveal that the recogni- tion rate of the SPLSR method can reach up to 79.23% and is superior to other compared dimensionality reduction methods.

This paper presents the classification of musical instruments using Mel Frequency Cepstral Coefficients (MFCC) and Higher Order Spectral features. MFCC, cepstral, temporal, spectral, and timbral features have been widely used in the task of musical instrument classification. As music sound signal is generated using non-linear dynamics, non-linearity and non-Gaussianity of the musical instruments are important features which have not been considered in the past. In this paper, hybridisation of MFCC and Higher Order Spectral (HOS) based features have been used in the task of musical instrument classification. HOS-based features have been used to provide instrument specific information such as non-Gaussianity and non-linearity of the musical instruments. The extracted features have been presented to Counter Propagation Neural Network (CPNN) to identify the instruments and their family. For experimentation, isolated sounds of 19 musical instruments have been used from McGill University Master Sample (MUMS) sound database. The proposed features show the significant improvement in the classification accuracy of the system.

This work examines the reduced-cost design optimization of dual- and multi-band antennas. The primary challenge is independent yet simultaneous control of the antenna responses at two or more frequency bands. In order to handle this task, a feature-based optimization approach is adopted where the design objectives are formulated on the basis of the coordinates of so-called characteristic points (or response features) of the antenna response. Due to only slightly nonlinear dependence of the feature points on antenna geometry parameters, optimization can be attained at a low computational cost. Our approach is demonstrated using two antenna structures with the optimum designs obtained in just a few dozen of EM simulations of the respective structure.

The task of generating fast and accurate three-dimensional (3D) models of objects or scenes through a sequence of non-calibrated images is an active field of research. The recent development in algorithm optimization has resulted in many automatic solutions that can provide an accurate 3D model from texture-full objects. Structure-from-motion (SfM) is an image-based method that uses discriminative point-based feature identifier, such as SIFT, to locate feature points in the images. This method faces difficulties when presented with the objects made of homogenous or texture-less surfaces. To reconstruct such surfaces a well-known technique is to apply an artificial variety by covering the surface with a random texture pattern prior to the image capturing process. In this work, we designed three series of image patterns which are tested based on the contrast and density ratio which increases from the first to the last pattern within the same series. The performance of the patterns is evaluated by reconstructing the surface of a texture-less object and comparing it with the true data. Using the best-found patterns from the experiments, a 3D model of a Moai statue is reconstructed. The experimental results demonstrate that the density and structure of a pattern highly affects the quality of the reconstruction.