Reliability Of Recognition Research Articles

Possibilistic Clustering-Promoting Semi-Supervised Learning for EEG-Based Emotion Recognition.

The purpose of the latest brain computer interface is to perform accurate emotion recognition through the customization of their recognizers to each subject. In the field of machine learning, graph-based semi-supervised learning (GSSL) has attracted more and more attention due to its intuitive and good learning performance for emotion recognition. However, the existing GSSL methods are sensitive or not robust enough to noise or outlier electroencephalogram (EEG)-based data since each individual subject may present noise or outlier EEG patterns in the same scenario. To address the problem, in this paper, we invent a Possibilistic Clustering-Promoting semi-supervised learning method for EEG-based Emotion Recognition. Specifically, it constrains each instance to have the same label membership value with its local weighted mean to improve the reliability of the recognition method. In addition, a regularization term about fuzzy entropy is introduced into the objective function, and the generalization ability of membership function is enhanced by increasing the amount of sample discrimination information, which improves the robustness of the method to noise and the outlier. A large number of experimental results on the three real datasets (i.e., DEAP, SEED, and SEED-IV) show that the proposed method improves the reliability and robustness of the EEG-based emotion recognition.

КЛАССИФИКАЦИЯ ЗОНЫ МАНЕВРИРОВАНИЯ РТК НА ОСНОВЕ ТАКТИЛЬНОЙ И ЗРИТЕЛЬНОЙ ИНФОРМАЦИИ

To increase the capabilities and expand possible applications of robotic systems for specialpurposes, switching from the currently being adopted remote control systems to semi-autonomoussystems that monitor operator actions and perform part of his functions is proposed. Moving toautonomous control systems capable of functioning in the "silent" mode, in shielded areas andbeyond the range of radio communications is proposed as a next step. Such intellectualization ofon-board control systems will allow to eliminate fundamental limitations and disadvantagescaused by the communication channel and ensures the implementation of robot group control. It isshown that the basis for robot autonomy increasing of on-board control systems, both for movementand tool control, is onboard environment model generation and determining the coordinatesof the control object. External environment model and current coordinates makes it possible toautomate the trajectory planning and movement, which ensures the autonomous functioning ofrobotic systems. The complex problem of environment segmentation according to geometric andground passability is considered, taking into account the characteristics of the chassis, the geometryof the relief and the supporting properties of the ground. The existing methods are describedand the results of experimental studies are given to solve the following main tasks: – classificationof the operation zone according to geometric passability based data from the onboard technicalvision system; – ground types recognition according to the integrated technical vision system; –using the apparatus of neural networks to improve the reliability of ground types recognition; –determination of ground reference characteristics by measuring the reactions of the chassis duringmovement. The promising directions for further research in integrating tactile and visual informationto improve the reliability of the classification of operation zone according to the complexcriterion of geometric and support passability are formulated.

Reliability of Trigger Point Evaluation in the Lower Leg Muscles.

Myofascial trigger point diagnosis is a clinical palpatory skill dependent on the patient's subjective response. The inter- and intra-rater reliability of trigger point physical evaluation in the lower leg muscles has rarely been reported. Previous reliability studies suffered from the Kappa paradox. To evaluate the inter- and intra-rater reliability of trigger point recognition in the lower leg muscles implying a specific method to overcome the first Kappa paradox. A reliability study with pre-second examiner exclusion to correct prevalence index. Physical therapy outpatient clinic, Beer-Sheva, Israel. In sum, 86 soldiers aged 18-30 referred for physical therapy with a diagnosis of musculoskeletal pain consented to take part in this study; 26 were excluded for lacking trigger points, leaving 60 subjects for analysis (31 women, 29 men). Both legs were evaluated, and the results were analyzed separately for symptomatic (N = 87) and asymptomatic legs (N = 31). Each subject was evaluated three times, twice by one examiner, and once by a second examiner. Dichotomous findings including palpable taut-band, tenderness, referred pain, and relevance of referred pain were recorded. Inter-rater reliability for active trigger points ranged from 0.49 to 0.75 (median: 0.52) and intra-rater reliability ranged from 0.41 to 0.84 (median: 0.65) and. For total trigger points intra-rater reliability ranged from 0.52 to 0.79 (median: 0.67), and inter-rater reliability ranged from 0.44 to 0.77 (median: 0.66). Physical examination is a reliable method of trigger point evaluation in lower leg muscles, and it can be used as a diagnostic method for trigger point evaluation.

Set-valued mode recognition-based Bayesian estimation for nonlinear stochastic systems with unknown sensor mode

This paper proposes the problem of joint state estimation and mode recognition for nonlinear stochastic systems with unknown sensor mode. The considered sensor mode is represented by a random finite set, whose elements can be one specific mode or a set of certain modes. A set-valued mode recognition-based Bayesian estimation framework is proposed to propagate the posterior density of the state conditioned on sensor modes and measurements, where the mode is recognized based on the maximum correntropy criterion. Furthermore, a mode-separability metric is proposed to discern the reliability of mode recognition, and utilized to derive two distinct implementation schemes, including state estimation based on separable and inseparable modes. Simulation results of fault detection and target tracking are provided to demonstrate the superiority of the proposed method in terms of state estimation accuracy and mode recognition effectiveness.

Development of an image segmentation model based on a convolutional neural network

This paper has considered a model of image segmentation using convolutional neural networks and studied the process efficiency based on models involving training the deep layers of convolutional neural networks. There are objective difficulties associated with determining the optimal characteristics of neural networks, so there is an issue related to retraining the neural network. Eliminating retraining by determining the optimal number of epochs only would not suffice since it does not provide high accuracy. The requirements for the set of images for training and model verification were defined. These requirements are best met by the image sets PASCAL VOC (United Kingdom) and NVIDIA-Aerial Drone (USA). It has been established that AlexNet (Canada) is a trained model and could perform image segmentation while object recognition reliability is insufficient. Therefore, there is a need to improve the efficiency of image segmentation. It is advisable to use the AlexNet architecture to build a specialized model, which, by changing the parameters and retraining some layers, would allow for a better process of image segmentation. Five models have been trained using the following parameters: learning speed, the number of epochs, optimization algorithm, the type of learning speed change, a gamma coefficient, a pre-trained model. A convolutional neural network has been developed to improve the accuracy and efficiency of image segmentation. Optimal neural network training parameters have been determined: learning speed is 0.0001, the number of epochs is 50, a gamma coefficient is 0.1, etc. An increase in accuracy by 3 % was achieved, which makes it possible to assert the correctness of the choice of the architecture for the developed network and the selection of parameters. That allows this network to be used for practical tasks related to image segmentation, in particular for devices with limited computing resources

Devising a method for recognizing the causes of deviations in the development of the plant Aloe arborescens L. using machine learning capabilities

This paper considers the process of developing a method to recognize the causes of plant growth deviations from normal using the advancements in artificial intelligence. The medicinal plant Aloe arborescens L. was chosen as the object of this research given that this plant had been for decades one of the best-selling new products in the world. Aloe arborescens L. is famous for its medicinal properties used in medicine, cosmetology, and even the food industry. Diagnosing the abnormalities in the plant development in a timely and accurate manner plays an important role in preventing the loss of crop production yields. The current study has built a method for recognizing the causes of abnormalities in the development of Aloe arborescens L. caused by a lack of watering or lighting, based on the use of transfer training of the VGG-16 convolutional neural network (United Kingdom). A given architecture is aimed at recognizing objects in images, which is the main reason for using it to achieve the goal set. The analysis of the quality metrics of the proposed image classification process by specified classes has revealed high recognition reliability (for a normally developing plant, 91 %; for a plant without proper watering, 89 %; and for a plant without proper lighting, 83 %). The analysis of the validity of test sample recognition has demonstrated a similar validity of the plant's classification to one of three classes: 92.6 %; 87.5 %; and 85.5 %, respectively. The results reported here make it possible to supplement the automated systems that control the mode parameters of hydroponic installations by the world's major producers with the main feedback on the deviation of the plant's development from the specified values

Development of an image segmentation model based on a convolutional neural network

This paper has considered a model of image segmentation using convolutional neural networks and studied the process efficiency based on models involving training the deep layers of convolutional neural networks. There are objective difficulties associated with determining the optimal characteristics of neural networks, so there is an issue related to retraining the neural network. Eliminating retraining by determining the optimal number of epochs only would not suffice since it does not provide high accuracy. The requirements for the set of images for training and model verification were defined. These requirements are best met by the image sets PASCAL VOC (United Kingdom) and NVIDIA-Aerial Drone (USA). It has been established that AlexNet (Canada) is a trained model and could perform image segmentation while object recognition reliability is insufficient. Therefore, there is a need to improve the efficiency of image segmentation. It is advisable to use the AlexNet architecture to build a specialized model, which, by changing the parameters and retraining some layers, would allow for a better process of image segmentation. Five models have been trained using the following parameters: learning speed, the number of epochs, optimization algorithm, the type of learning speed change, a gamma coefficient, a pre-trained model. A convolutional neural network has been developed to improve the accuracy and efficiency of image segmentation. Optimal neural network training parameters have been determined: learning speed is 0.0001, the number of epochs is 50, a gamma coefficient is 0.1, etc. An increase in accuracy by 3 % was achieved, which makes it possible to assert the correctness of the choice of the architecture for the developed network and the selection of parameters. That allows this network to be used for practical tasks related to image segmentation, in particular for devices with limited computing resources

Augmented reality visualization tool for the future of tactical combat casualty care.

The objective of this project was to identify and develop software for an augmented reality application that runs on the US Army Integrated Visual Augmentation System (IVAS) to support a medical caregiver during tactical combat casualty care scenarios. In this augmented reality tactical combat casualty care application, human anatomy of individual soldiers obtained predeployment is superimposed on the view of an injured war fighter through the IVAS. This offers insight into the anatomy of the injured war fighter to advance treatment in austere environments.In this article, we describe various software components required for an augmented reality tactical combat casualty care tool. These include a body pose tracking system to track the patient's body pose, a virtual rendering of a human anatomy avatar, speech input to control the application and rendering techniques to visualize the virtual anatomy, and treatment information on the augmented reality display. We then implemented speech commands and visualization for four common medical scenarios including injury of a limb, a blast to the pelvis, cricothyrotomy, and a pneumothorax on the Microsoft HoloLens 1 (Microsoft, Redmond, WA).The software is designed for a forward surgical care tool on the US Army IVAS, with the intention to provide the medical caregiver with a unique ability to quickly assess affected internal anatomy. The current software components still had some limitations with respect to speech recognition reliability during noise and body pose tracking. These will likely be improved with the improved hardware of the IVAS, which is based on a modified HoloLens 2.

A comparative study on handcrafted features v/s deep features for open-set fingerprint liveness detection

Abstract A fingerprint liveness detector is a pattern classifier that is used to distinguish a live finger from a fake (spoof) one in the context of an automated fingerprint recognition system. As liveness detectors or presentation attack detectors are vulnerable to presentation attacks, the security and reliability of fingerprint recognition are compromised. Presentation attack detection mechanisms rely on handcrafted or deep features to classify an image as live or spoof. In addition, to strengthen the security, fingerprint liveness detectors should be robust to presentation attacks fabricated using unknown fabrication materials or fingerprint sensors. In this paper, we conduct a comprehensive study on the impact of handcrafted and deep features from fingerprint images on the classification error rate of the fingerprint liveness detection task. We use LBP, LPQ and BSIF as handcrafted features and VGG-19 and Residual CNN as deep feature extractors for this study. As the problem is targeted as an open-set problem, the emphasis is on achieving better robustness and generalization capability. In our observation, handcrafted features outperformed their deep counterparts in two of the three cases under the within-dataset environment. In the cross-sensor environment, deep features obtained a better accuracy, and in the cross-dataset environment, handcrafted features obtained a lower classification error rate.

Reliability of landmark recognition and angle measurement of lower extremity X-ray using deep learning algorithm

Purpose: Analyzing X-ray is an essential process for the diagnosis and treatment planning of patients with knee osteoarthritis. However, it is a repetitive, energy consuming work in a daily practice. A deep learning algorithm can recognize specific anatomical landmark of lower extremity and can analyze various parameters. This study hypothesized that there would be a significant difference in measurement between human and algorithm. The object of this study is to develop and validate a deep learning algorithm that can automatically recognize and analyze lower extremity X-ray.

METHOD OF IDENTIFICATION OF ELEMENTS OF A COMPLEX SYSTEM

The article proposes a method of using identification in automated systems that provide identification of elements of various complex systems that are in variable conditions. The existing methods do not take into account the set of variable conditions that affect the characteristics of the identified elements of a complex system, while reducing the number of favorable outcomes in the process of identifying elements in the composition of electronic means. The proposed method makes it possible to increase the reliability of identification of elements of a complex system that are in various variable conditions by creating standards and implementations that correspond to the conditions in which the identified elements of a complex system are located. A formalized algorithm for implementing the methodology is presented. The distinctive features of the developed method are the automatic determination of the reliability of recognition of elements of a complex system in non-stationary conditions.

Reliability of Bystander Recognition of Clinical Features in Pre-Hospital Classification of Acute Cerebrovascular Syndromes: Preliminary Findings.

The recognition of stroke symptoms by patients or bystanders directly affects the outcomes of patients with acute cerebrovascular disease. The objective of the study was to assess the predictive value of the medical his- tory and clinical features recognized by the patients' bystanders to classify neurovascular syndromes in pre-hospital settings. We included 150 stroke patients of two Mexican referral centers: 50 with acute ischemic stroke (AIS), 50 with intracerebral hemorrhage (ICH), and 50 with subarachnoid hemorrhage (SAH). The performance of clinical prediction rules (CPR) to identify the stroke types was evaluated with features recognized by the patients' bystanders before hospital arrival. The impact of CPRs on early arrival and in-hospital mortality was also analyzed. Overall, 72% of the patients had previous medical evaluations in other centers before final referral to our hospitals, and therefore, only 45% had a final onset- to-door time <6 h, even when the first medical assessment had occurred in ≤1 h in 75% of cases. Clinical features noticed by the patients' bystanders had low positive predictive values (PPV) for any stroke type. The CPR "language or speech disor- der + focal motor deficit" had 93% sensitivity and a negative predictive value (NPV) of 84% to distinguish AIS. In SAH, head- ache alone showed a sensitivity of 84% and NPV of 97%. No CPR had an adequate performance on ICH. CPRs were not as- sociated with final onset-to-door time. Altered consciousness, age ≥65 years, indirect arrival with stops before final referral, and atrial fibrillation increased in-hospital mortality. Clinical features referred by the witness of a neurovascular emergency have limited PPV, but adequate NPV in ruling-out AIS and SAH among stroke types. The use of CPRs had no impact on onset-to-door time or in-hospital mortality when the final arrival to a third-level center occurs with previous medical refer- rals.

A Study on Face Recognition and Reliability Improvement Using Classification Analysis Technique

In this study, we try to find ways to recognize face recognition more stably and to improve the effectiveness and reliability of face recognition. In order to improve the face recognition rate, a lot of data must be used, but that does not necessarily mean that the recognition rate is improved. Another criterion for improving the recognition rate can be seen that the top/bottom of the recognition rate is determined depending on how accurately or precisely the degree of classification of the data to be used is made. There are various methods for classification analysis, but in this study, classification analysis is performed using a support vector machine (SVM). In this study, feature information is extracted using a normalized image with rotation information, and then projected onto the eigenspace to investigate the relationship between the feature values through the classification analysis of SVM. Verification through classification analysis can improve the effectiveness and reliability of various recognition fields such as object recognition as well as face recognition, and will be of great help in improving recognition rates.

ВЕРИФІКАЦІЯ ТА РОЗПІЗНАВАННЯ ПІДПИСУ ЯК БАГАТОПАРАМЕТРИЧНОГО ПРОЦЕСУ НА ОСНОВІ СПАЙКІНГОВОЇ НЕЙРОННОЇ МЕРЕЖІ

The article reviews the known methods of dynamic signature verification, which are summarized in the classification table. A method of dynamic signature verification based on a spiking neural network is proposed. Three dynamic parameters of the signature l(t), Dα(t), Z(t) &nbsp;are chosen, which are invariant to the angle of inclination of the signature, and after their normalization - also to the spatial and temporal scales of the signature. These dynamic signature parameters are simultaneously fed to the spiking neural network for recognition in the form of time series without prior conversion into a vector of static features. This, on the one hand, simplifies the method due to the absence of complex computational conversion procedures and, on the other hand, prevents the loss of useful information and therefore increases the accuracy and reliability of signature verification and recognition (especially for forgery signatures that are highly correlated with genuine ones). The used neural network has a simple learning procedure, and not all neurons of the network are trained, but only the output neurons. If you need to add new signatures, you do not need to retrain the entire network, but just add a few output neurons and learn only their connections.

Validity and reliability of tooth color selection by smartphone photography and software applications.

Aim:This study assessed the validity and reliability of color selection by smartphone photography using two smartphone applications and Adobe Photoshop software.Settings and Design:In vitro comparative study.Materials and Methods:The validity and reliability of dental tooth shade recognition (DTSR), Chromatcher, and Adobe Photoshop were evaluated for color selection of shade tabs. The iPhone 7 camera in automatic mode was used for photography. Images were captured using Smile Lite with/without polarized filter and with camera flash. To assess the reliability, nine Vita Lumin Vacuum shade tabs were chosen and each was photographed for 10 times using Smile Lite. The reliability of DTSR, Chromatcher, and Photoshop in shade-taking was calculated. To assess their validity, 16 shade tabs of Vita Lumin Vacuum and 26 shade tabs of Vita 3D Master were photographed using the aforementioned lighting conditions. The color of photographs was calibrated and shade-taking was performed and compared with the shade suggested by SpectroShade as reference.Statistical Analysis Used:Data were analyzed using Two-way analysis of variance and Bonferroni post hoc test.Results:The reliability of Photoshop, DTSR, and Chromatcher was 98.88%, 63.3%, and 100%, respectively. The validity of Photoshop was significantly higher than other software programs (P < 0.05). Chromatcher had higher validity than DTSR (P < 0.05).Conclusion:Shade-taking by calibrated smartphone pictures and Adobe Photoshop has high validity and reliability.

РОЗРОБЛЕННЯ БІОМЕТРИЧНИХ МЕТОДІВ ІДЕНТИФІКАЦІЇ НА ПІДСТАВІ ФІЛЬТРАЦІЇ ATEB-ГАБОРОМ

The article is devoted to the development of biometric identification methods based on new filtration methods. Biometric identification systems need constant improvement, because they often work slowly and give the wrong result. To increase the reliability of biometric image recognition, the method is formed, which is formed from the stages: segmentation, normalization, local orientation estimation, local estimation, spine frequency estimation, Gabor filter implementation, binarization, thinning. A new filtering method is proposed, which is based on a new type of function – Ateb-functions, which are used next to the Gabor filter. The local orientation can be calculated from local gradients using the arctangent function. The normalization process is performed to evenly redistribute the values ​​of image intensity. When segmenting, the foreground areas in the image are separated from the background areas. A new method of wavelet conversion of biometric image filtering based on Ateb-Gabor has been developed. The Gabor filter is used for linear filtering and improves the quality of the converted image. Symmetry and wavelet transform operations are also used to reduce the number of required multiplication and addition operations. The method is based on the well-known Gabor filter and allows you to rearrange the image with clearer contours. Therefore, this method is applicable to biometric images, where the creation of clear contours is particularly relevant. When Gabor filtering, the image is reconstructed by multiplying the harmonic function by the Gaussian function. Ateb functions are a generalization of elementary trigonometry, and, accordingly, have greater functionality. Ateb-Gabor filtering allows you to change the intensity of the whole image, as well as the intensity in certain ranges, and thus make certain areas of the image more contrasting. Filtering with Ateb functions allows you to change the image from two rational parameters. This allows you to more flexibly manage filtering and choose the best options. When you perform a thinning, the foreground pixels are erased until there is one pixel wide. A standard thinning algorithm is used, or the thinning developed by the authors in other studies. This filtering will provide more accurate characteristics, as it allows you to get more sloping shapes and allows you to organize a wider range of curves. Numerous experimental studies indicate the effectiveness of the proposed method.

On the Generalization and Reliability of Single Radar-Based Human Activity Recognition

Identifying human activities using short-range and low-power radars has attracted much attention among the researchers and consumer electronics industry. This paper considers human activity recognition in the context of a single Frequency Modulated Continuous Wave (FMCW) radar as the measurement tool. A classification pipeline is proposed to handle the data pre-processing and feature extraction and a machine-learning based solution is devised to undertake the activity classification. The performance of the proposed architecture is evaluated under both unseen subjects and new room layouts. We show how the accuracy of the activity classification will be affected by situations such as poor aspect-angle and occlusions created by furniture that normally arise in realistic scenarios where an unseen layout is considered. A two-stage classifier will be then proposed to enhance the generalization of the model, especially, to unseen rooms. Besides, an extensive feature exploration will be conducted and the importance of features in the generalization will be studied. The results in this paper will conclude a machine learning pipeline that will generalize well to unseen subjects and new room layouts, which are two main difficulties that arise in most radar-based activity classification tasks.

Искусственный интеллект в задачах распознавания динамических жестов

The development of effective methods for recognizing human hand gestures is an urgent task, both from a scientific and an applied point of view. Gesture recognition methods are the basis of contactless interfaces for managing technical systems. The most significant areas of application of gesture recognition systems are automatic sign language translation and contactless control of technical systems. Almost all existing implementations of recognition methods work confidently when the hand is on a uniform background. But in reality, such a case of using such implementations is unlikely. The likelihood of correct recognition drops sharply against a complex background when the hand in the image crosses the face or body parts with exposed skin. Additional restrictions on algorithms impose requirements for their hardware implementation. These should be compact devices on a chip (SoC) with low consumption, size and price, and, therefore, with low computing power. Taking into account the above, it should be recognized that it is relevant to solve the problem of increasing the efficiency of algorithms and methods of gesture recognition. Two approaches to solving the problem of recognizing dynamic hand gestures are considered - analytical and neural network. It is shown that the use of artificial intelligence techniques can increase the reliability of gesture recognition in complex video surveillance conditions. However, the use of neural network algorithms cannot show high performance on single-board computers if they do not have an NPU or a powerful GPU module.

Support subspace method and conforming estimation in autonomous navigation tasks

The article discusses the technology of autonomous navigation by optical observations. A scheme based on the recognition of three-dimensional landmarks is considered. To improve the accuracy and reliability of recognition, a survey model is constructed for each landmark. When calculating the parameters of this model, the most conformed data is selected. The problem of recognition is solved using the method of support subspaces. Support subspaces are formed from a set of vectors of each reference landmark. An important advantage of the technology is the ability to refine the landmark models based on the results of current observations during operation.

A staked discriminative auto-encoder based on center loss for radar target HRRP recognition

In the radar automatic target recognition filed, extracting representative features from the radar high resolution range profile (HRRP) is the key issues, which determines the accuracy and reliability of radar target recognition. In this paper, we propose a novel stacked discriminative auto-encoder(S-disAE), the center loss is integrated into auto-encoder, it can force the learned feature with large distance close to their class feature representation center, so as to reduce the intra-class variations while keeping the features of different classes separable. In pre-training stage of our discriminative auto-encoder(disAE), we combine the mean square error loss and center loss to learn the main factors of input raw data and minimize the intra-class variations of the feature, in fine-tuning stage, we also combine the softmax cross entropy loss and center loss to improve the classification accuracy. We conduct several experiments on the simulated HRRP data, the results demonstrate that the proposed algorithm can extract discriminant features and improve the HRRP recognition accuracy.