Search

The main scientific results of the 16th International Conference on Pattern Recognition and Information Processing (PRIP-2023), Minsk, Republic of Belarus, October 2023, are reviewed and analyzed. The history of this series of conferences is outlined, and its significant role in the development of the theory and practice of image analysis, pattern recognition, and artificial intelligence is indicated. A list of articles in the special issue is provided, prepared from reports selected by the PRIP-2023 Program Committee. [ABSTRACT FROM AUTHOR]

The implementation of Deep Learning (DL) Prediction techniques for Human Age Prediction (HAP) has been widely researched and studied to prevent, treat, and extend life expectancy. While most algorithms rely on facial images, MRI scans, and DNA methylation for training and testing, they are seldom implemented due to a lack of significant validation and evaluation in real-world scenarios, low performance, and technical challenges. To address these issues, this paper proposes the Data, Classification Technique, Prediction, and View (DCPV) taxonomy, which outlines the primary components required to implement and validate a deep learning model for predicting human age. By providing a common baseline for end-users and researchers, this taxonomy offers a clearer view of the constituents of deep learning prediction approaches, enabling the development of similar systems in the health domain. In contrast to existing machine learning methods, the proposed taxonomy emphasizes the value of deep learning practices based on performance, accuracy, and efficiency in predicting human age. To validate the DCPV taxonomy, the study examines 31 state-of-the-art research journal articles within the HAP system domain, assessing the taxonomy's performance, accuracy, robustness, and model comparisons. [ABSTRACT FROM AUTHOR]

Patterns are crucial for efficiently scheduling microservice workflow applications to containers in cloud computing scenarios. However, it is challenging to learn patterns of microservice workflows because of their complex precedence constrained structures provided by users with more lightweighted, diversified, and personalized services. In this paper, we propose a graph neural network is designed to identify patterns within a set of microservice workflows by mining the common substructures of workflows. Based on the learned patterns, a pattern-based scheduling algorithm framework is developed for microservice workflows with soft deadline constraints to minimize the average tardiness. A sorting strategy is introduced based on urgency and pattern coverage rate. For simplification of the task sorting process, the pattern-based task sorting algorithm (PB-TS) is devised. Furthermore, a resource selection phase is incorporated to the pattern-based resource selection algorithm (PB-RS) to minimize the candidate resource space. Experimental results demonstrate the proposed method is much efficient as compared to three classical algorithms. [ABSTRACT FROM AUTHOR]

Belief and plausibility functions based on evidence theory (ET) have been widely used in managing uncertainty. Various generalizations of ET to fuzzy sets (FSs) have been reported in the literature, but no generalization of ET to q-rung orthopair fuzzy sets (q-ROFSs) has been made yet. Therefore, this paper proposes a novel, simple, and intuitive approach to distance and similarity measures for q-ROFSs based on belief and plausibility functions within the framework of ET. This research addresses a significant research gap by introducing a comprehensive framework for handling uncertainty in q-ROFSs using ET. Furthermore, it acknowledges the limitations inherent in the current state of research, notably the absence of generalizations of ET to q-ROFSs and the challenges in extending belief and plausibility measures to certain aggregation operators and other generalizations including Hesitant fuzzy sets, Bipolar fuzzy sets, Fuzzy soft sets etc. Our contribution lies in the proposal of a novel approach to distance and similarity measures for q-ROFSs under ET, utilizing Orthopairian belief and plausibility intervals (OBPIs). We establish new similarity measures within the generalized ET framework and demonstrate the reasonability of our method through useful numerical examples. Additionally, we construct Orthopairian belief and plausibility GRA (OBP-GRA) for managing daily life complex issues, particularly in multicriteria decision-making scenarios. Numerical simulations and results confirm the usability and practical applicability of our proposed method in the framework of ET. [ABSTRACT FROM AUTHOR]

Due to the digitization of data and the dynamic requirements of the users, cloud computing is one of the most used technologies in the present scenario. Cloud computing provides a platform to store, process, and share data remotely for heterogeneous users and provides services according to the requests generated by those users. However, its rapid growth has led to one of the major challenges in the environment; the security and privacy of the data. To address the security and privacy concerns, in this paper, our major contribution is a trusted computing framework namely Secure Framework using Behavior and Role Analysis (SFBRA) for cloud data security. The framework utilizes user log monitoring data, pattern recognition algorithms, and role-based access mechanisms to detect malicious and suspicious activities of different users. Our proposed framework provides two levels of security for cloud users. In Level-1, we calculate the trust value of the logged-in users by analyzing the existing log table and pattern of request access. In Level-2, we calculate the trust of the request (storage, processing, sharing) data packet using behavior analysis of the user and a role-based access mechanism and finally detect the malicious activities. The efficacy of our proposed framework is demonstrated through experimentation, where we compare our framework with existing research works. The results show 95% accuracy in potential attack detection and prevention,approximately 8 Mbps throughput, and 0.003% packet drop on average. [ABSTRACT FROM AUTHOR]

In this paper proposes an algorithm for the analytical calculation of convolutional neural networks without using neural network training algorithms. A description of the algorithm is given, on the basis of which the weights and threshold values of a convolutional neural network are analytically calculated. In this case, to calculate the parameters of the convolutional neural network, only 10 selected samples were used from the MNIST digit database, each of which is an image of one of the recognizable classes of digits from 0 to 9, and was randomly selected from the MNIST digit database. As a result of the operation of this algorithm, the number of channels of the convolutional neural network layers is also determined analytically. Based on the proposed algorithm, a software module was implemented in the Builder environment C++, on the basis of which a number of experiments were carried out with recognition of the MNIST database. The results of the experiments described in the work showed that the computation time of convolutional neural networks is very short and amounts to fractions of a second or a minute. Analytically computed convolutional neural networks were tested on the MNIST digit database, consisting of 1000 images of handwritten digits. The experimental results showed that already using only 10 selected images from the MNIST database, analytically calculated convolutional neural networks are able to recognize more than half of the images of the MNIST database, without application of neural network training algorithms. In general, the study showed that artificial neural networks, and in particular convolutional neural networks, are capable of not only being trained by learning algorithms, but also recognizing images almost instantly, without the use of learning algorithms using preliminary analytical calculation of the values of the neural network's weights. [ABSTRACT FROM AUTHOR]

Automatic emotion recognition from speech signals is an important research area. Many speech emotion recognition (SER) methods have been proposed, among which ensemble learning is an effective way to recognize speech emotion. However, the ability and diversity of the base classifier are not carefully considered in the case of limited available speech emotion samples. To overcome the above problem, this paper proposes a new diversity subspace generation based on feature selection (DSGFS) for SER. In DSGFS, a constrained problem is cleverly designed, which can iteratively select many diversity and strong subspaces, so the classification ability and the diversity of the corresponding base classifiers can be ensured. As a result, more features can be extracted from the data, an ensemble classifier framework with strong base classifiers can be automatically generated, and the number of base classifiers can be smaller. The proposed models offered SER weighted average recall of 87.24%, 64.58%, 69.10%, 53.50% on the EmoDB, SAVEE, RAVDESS, CASIA datasets with speaker independent, respectively, which validate the proposed approach in terms of the performance of speech emotion recognition. [ABSTRACT FROM AUTHOR]

The use of a convolutional neural network with transfer learning is a strategy that defines high-level features, commonly explored to study patterns in medical images. These features can be analyzed via different methods in order to design hybrid models with more useful and accurate solutions for clinical practice. In this paper, a computational scheme is presented to define hybrid models through deep features by transfer learning, selection by ranking and a robust ensemble classifier with five algorithms. The obtained models were applied to classify histological images from breast, colorectal and liver tissue. The strategy developed here allows knowing important results and conditions to improve models of computer-aided diagnosis, even exploring classic CNN models. The features were defined using layers from the AlexNet and ResNet-50 architectures. The attributes were organized into subsets of the most relevant features and submitted to a k-fold cross-validation process. The best hybrid models were obtained with deep features from the ResNet-50 network, using distinct layers (activation_48_relu and avg_pool) and a maximum of 35 descriptors. These hybrid models provided 98.00% and 99.32% of accuracy values, with emphasis on histological images of breast cancer, indicating the best solution among those available in the specialized Literature. Also, these models provided more relevant results for classifying UCSB and LG datasets than regularized techniques and CNN architectures, exploring data augmentation or not. The computational scheme with detailed information regarding the main hybrid models is a relevant contribution to the community interested in the study of machine learning techniques for pattern recognition. [ABSTRACT FROM AUTHOR]

In recent years, Deep Neural Networks (DNN) have emerged as a practical method for image recognition. The raw data, which contain sensitive information, are generally exploited within the training process. However, when the training process is outsourced to a third-party organization, the raw data should be desensitized before being transferred to protect sensitive information. Although masks are widely applied to hide important sensitive information, preventing inpainting masked images is critical, which may restore the sensitive information. The corresponding models should be adjusted for the masked images to reduce the degradation of the performance for recognition or classification tasks due to the desensitization of images. In this paper, we propose a mask-based image desensitization approach while supporting recognition. This approach consists of a mask generation algorithm and a model adjustment method. We propose exploiting an interpretation algorithm to maintain critical information for the recognition task in the mask generation algorithm. In addition, we propose a feature selection masknet as the model adjustment method to improve the performance based on the masked images. Extensive experimentation results based on multiple image datasets reveal significant advantages (up to 9.34% in terms of accuracy) of our approach for image desensitization while supporting recognition. [ABSTRACT FROM AUTHOR]

Picture fuzzy sets (PFSs), as a generalization of traditional fuzzy sets and intuitionistic fuzzy sets (IFSs), offer a powerful framework for modeling and dealing with imprecise and uncertain information and have been widely used in various fields. Nevertheless, how to effectively measure the differences or similarities between PFSs is still a challenging and urgent problem. Despite that some distance measures for PFSs have been developed, some of them have counterintuitive and unreasonable results and even partially fail to satisfy the axiomatic definition of distance measures. To handle these limitations, in this paper, we propose two distance measures of PFSs motivated by the merits of Jensen–Shannon divergence. Specifically, the first distance measure of the PFSs considers the positive degree, neutral degree and negative degree. The second distance measure introduces the refusal degree on this basis. Moreover, the proposed distance measures not only satisfy the axiomatic definition but also overcome the counterintuitive results of existing distance measures. Finally, two kinds of applications related to pattern recognition and medical diagnosis are used to verify the performance of the proposed distance measures. [ABSTRACT FROM AUTHOR]

Divergence measure between intuitionistic fuzzy sets (IFSs) is important due to its wide range of applications in various fields including pattern recognition, image segmentation, decision-making and clustering. This paper introduces the characterization of belief and plausible intuitionistic fuzzy sets (BP-IFSs) to explore novel logarithmic and non-logarithmic divergence measures between two BP-IFSs. These measures are regarded as highly useful approaches to express ambiguous information within the framework of Dempster–Shafer Theory (DST). An axiomatic definition based on proposed divergence measures is also stated within a frame work of newly established theory. Furthermore, the proposed divergence measures are utilized in three different applications: (i) an example related to the recognition of BP-IFS patterns is provided to demonstrate the practicality of the proposed method in pattern recognition. (ii) An example of Hierarchical agglomerative clustering is also provided. (iii) Introduces an innovative Belief and Plausible Combinative Distance-based Assessment (BP-CODAS) method based on proposed measures for resolving Multicriteria Decision Making (MCDM) problems connected to child labor in under developed countries. The examples provided in these different directions are sufficient to demonstrate the effectiveness, applicability and viability of the suggested methods within the framework of generalized DST. [ABSTRACT FROM AUTHOR]

Abstract—On February 6, 2023, two devastating earthquakes struck southern central Turkey, nine hours apart. The ground shaking from these earthquakes even swept over a significant part of northwestern Syria. In this paper, we consider the locations of the epicenters of these earthquakes against the prediction of the М ≥ 6.5 earthquake epicenters in Anatolia and adjacent regions, which was made in 1973 by I.M. Gelfand, V.I. Keilis-Borok and their colleagues using the Kora-3 pattern recognition algorithm based on morphostructural zoning data. We also present the results of early detection of the periods with an increased probability of the strongest earthquakes as determined by the medium-term prediction algorithm for the M8 earthquakes. It is found that the epicenters of the February 6, 2023 earthquakes occurred in a node identified in 1973 as potentially earthquake-prone for М ≥ 6.5, and both events occurred within the spatial and temporal coverage of the warning area diagnosed by the M8 algorithm in July 2021. [ABSTRACT FROM AUTHOR]

The prevalence of skin illnesses is higher than that of other diseases. Fungal infection, bacteria, allergies, viruses, genetic factors, and environmental factors are among important causative factors that have continuously escalated the degree and incidence of skin diseases. Medical technology based on lasers and photonics has made it possible to identify skin illnesses considerably more rapidly and correctly. However, the cost of such a diagnosis is currently limited and prohibitively high and restricted to developed areas. The present paper develops a holistic, critical, and important skin disease prediction system that utilizes machine learning and deep learning algorithms to accurately identify up to 20 different skin diseases with a high F1 score and efficiency. Deep learning algorithms like Xception, Inception-v3, Resnet50, DenseNet121, and Inception-ResNet-v2 were employed to accurately classify diseases based on the images. The training and testing have been performed on an enlarged dataset, and classification was performed for 20 diseases. The algorithm developed was free from any inherent bias and treated all classes equally. The present model, which was trained using the Xception algorithm, is highly efficient and accurate for 20 different skin conditions, with a dataset of over 10,000 photos. The developed system was able to classify 20 different dermatological diseases with high accuracy and precision. [ABSTRACT FROM AUTHOR]

QR codes are often placed on complex backgrounds and under unstable illumination conditions, which in fact renders their localization and decoding quite challenging. In this paper, we propose a robust system devised to ensure two main tasks: firstly, extracting QR codes from captured images regardless of their quality. Secondly, enhance the extracted QR codes textures so that they can be decoded correctly. A QR code is principally localized through three finder patterns (FPs), which are placed in its three corners, as well as a set of alignment patterns (APs). FP and AP are characterized by specific textures (alternative black and white pixel sequences), defined respectively by the following ratios: 1:1:3:1:1 and 1:1:1:1:1. The proposed system starts by segmenting the captured image using an achromatic filter. This is followed by horizontal and vertical scans, aiming at extracting the valid horizontal and vertical segments satisfying the aforementioned ratios. Afterwards, the valid segment intersection is computed to localize candidate areas containing FPs and APs. The rough extracted patterns are filtered based on Multiclass Support Vector Machine classifier combined with the histogram of Oriented Gradients. The retained patterns are then transmitted to a pattern closeness measurement algorithm to conveniently localize QR code positions. In order to evaluate the robustness and weakness of the proposed approach, a comparative study has been carried out. In this study, we compared our approach against the most relevant competitive QR code extractors, such as Zxing, Zbar, Quirc, Leadtools and Dynamsoft. Our algorithm outperforms the evoked extractors with a precision rate of 97,3% and 96% under motion blur and JPEG compression respectively. Furthermore, the execution time analysis proved that our algorithm is not computationally expensive with a maximum running time of 1092 ms at a high resolution image and high number of QR codes per image.. [ABSTRACT FROM AUTHOR]

Fungi are one of the major causes that contributed to plant diseases. There are lots of fungi species but it is estimated that only 10% have been described. There are two major approaches to identifying fungi species, morphological identification, and molecular test which need cautious clarification to make good interpretations and are time-consuming. In this paper, we propose a Machine Learning approach that involves the use of the K-Means clustering technique, and Decision Tree to highlight the observed fungi spore images taken under the microscopic view and discard background pixels to produce digital images database which later can be used for Deep Learning. [ABSTRACT FROM AUTHOR]

Although there are many researches on the similarity measure for intuitionistic fuzzy sets, some of them have some defects. To overcome these shortcomings, based on the Euclidean distance between intuitionistic fuzzy set and three special intuitionistic fuzzy sets, a new similarity measure for intuitionistic fuzzy sets is proposed in this paper. The proposed similarity measure is compared with the existing similarity measures in intuitionistic fuzzy environment by numerical examples. Moreover, we apply the proposed similarity measure to pattern recognition problems in intuitionistic fuzzy environment. The experimental results show that the proposed similarity measure for intuitionistic fuzzy sets is reasonable and effective. [ABSTRACT FROM AUTHOR]

Machine Learning (ML) and Deep Learning (DL) have achieved high success in many textual, auditory, medical imaging, and visual recognition patterns. Concerning the importance of ML/DL in recognizing patterns due to its high accuracy, many researchers argued for many solutions for improving pattern recognition performance using ML/DL methods. Due to the importance of the required intelligent pattern recognition of machines needed in image processing and the outstanding role of big data in generating state-of-the-art modern and classical approaches to pattern recognition, we conducted a thorough Systematic Literature Review (SLR) about DL approaches for big data pattern recognition. Therefore, we have discussed different research issues and possible paths in which the abovementioned techniques might help materialize the pattern recognition notion. Similarly, we have classified 60 of the most cutting-edge articles put forward pattern recognition issues into ten categories based on the DL/ML method used: Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Generative Adversarial Network (GAN), Autoencoder (AE), Ensemble Learning (EL), Reinforcement Learning (RL), Random Forest (RF), Multilayer Perception (MLP), Long-Short Term Memory (LSTM), and hybrid methods. SLR method has been used to investigate each one in terms of influential properties such as the main idea, advantages, disadvantages, strategies, simulation environment, datasets, and security issues. The results indicate most of the articles were published in 2021. Moreover, some important parameters such as accuracy, adaptability, fault tolerance, security, scalability, and flexibility were involved in these investigations. [ABSTRACT FROM AUTHOR]

Surface roughness is a major factor controlling the shear strength of discontinuities. Various methods have so far been presented for determining surface roughness based on 2D profiles. These methods use statistical, fractal, geostatistical, directional, and spectral features of the surface profiles most of which were established through the relationship between the features and Barton standard profiles. To overcome some of the shortcomings of the standard JRC profiles, this paper presents a new method based on the investigation of 2D profiles within “pattern recognition” framework for rock surface recognition. In this way, more than 9000 profiles were gathered from 84 natural rock samples for the calculation of the features, and the joint roughness coefficient was back-calculated from test results. A representative feature vector and profile of a surface were defined, and two methods of principal component analysis (PCA) and liner discriminant analysis (LDA) were applied in order to prepare inputs for the rock surface classification. Then, minimum mean distance (MMD) and K-nearest neighbors (KNN) were used for the classification. The results show that the latter provides lower classification error (44.4% when a level difference between the predicted and back-calculated classes is accepted) due to reducing the effect of dissimilar samples. Comparison of the results with those of single-variable equations that are based on statistical (δ and Z 2 ) and directional parameters θ max ∗ 1 + C with 53.57%, 52.38%, and 48.8% errors, respectively, proves that the proposed procedure is an efficient tool for the estimation of JRC. Furthermore, based on the results obtained from the pattern recognition analysis, 10 natural and artificial reference profiles attained based on a fully quantified approach were proposed as a replacement for the current JRC profiles. [ABSTRACT FROM AUTHOR]

Mineral grain recognition is an extremely important task in many fields, especially in mineral exploration, when trying to identify locations where precious minerals can possibly be found. The usual manual method would be to collect samples; a specialized individual using expensive equipment would manually identify and then count the grain minerals in the sample. This is a tedious task that is time-consuming and expensive. It is also limited because small portions of areas can be surveyed; even then, it might require extremely long periods. In addition, this process is still prone to human errors. Developing an automatic system to identify, recognize, and count grain minerals in samples from images would allow for more precise results than the time required by humans. In addition, such systems can be fitted on robots that collect samples, take images of the samples, and then proceed with the automated recognition and counting algorithm without human intervention. Vast amounts of land can be surveyed in this way. This paper proposes a modified approach for microscopic grain mineral recognition and classification using hybrid features and ensemble algorithms from images. The enhanced approach also included a modified segmentation approach, which enhanced the results. For 10 classes of microscopic mineral grains, using the modified approach and the ensemble algorithm resulted in an average accuracy of 84.01%. For 8 classes, the average reported accuracy is 94.93% using the Boosting ensemble learning with the C4.5 classifier. The results obtained outperform similar methods reported in the extant literature. [ABSTRACT FROM AUTHOR]

Ensuring the durability of high-voltage (HV) cross-linked polyethylene (XLPE) cable insulation requires vigilant cleanliness maintenance during production to mitigate impurities, including oxidized parts and voids, which can compromise insulation integrity. Hence, this paper presents a MATLAB/Simulink partial discharge (PD) capacitive model of five well-known PD defects: crack, contamination, air void, microcrack, and composite, found in insulation materials HV XLPE insulation. Furthermore, this work proposes an extraordinary deep learning approach utilizing short-time Fourier transform (STFT) scalograms to represent PD signals in the time-frequency domain and train a convolutional neural network (CNN) to classify different PD defects. We focused on vital factors affecting STFT + CNN-aided pattern recognition accuracy, such as the number of network layers, convolutional kernel size, activation function, and pooling technique to optimize the network. Our study demonstrates that the proposed STFT + CNN approach outperforms traditional methods in recognizing PD patterns, especially for high signal similarity. Simulation results indicate that the STFT + CNN model achieves the highest classification accuracy of 0.9744 with minimal computation time (20 msec), making it suitable for real-time PD activity classification. [ABSTRACT FROM AUTHOR]

Highlights: The types, working principles, advantages and limitations of pattern recognition methods based on chemiresistive gas sensor array are reviewed and discussed comprehensively. Outstanding and novel advancements in the application of machine learning methods for gas recognition in different important areas are compared, summarized and evaluated. The current challenges and future prospects of machine learning methods in artificial olfactory systems are discussed and justified. As information acquisition terminals for artificial olfaction, chemiresistive gas sensors are often troubled by their cross-sensitivity, and reducing their cross-response to ambient gases has always been a difficult and important point in the gas sensing area. Pattern recognition based on sensor array is the most conspicuous way to overcome the cross-sensitivity of gas sensors. It is crucial to choose an appropriate pattern recognition method for enhancing data analysis, reducing errors and improving system reliability, obtaining better classification or gas concentration prediction results. In this review, we analyze the sensing mechanism of cross-sensitivity for chemiresistive gas sensors. We further examine the types, working principles, characteristics, and applicable gas detection range of pattern recognition algorithms utilized in gas-sensing arrays. Additionally, we report, summarize, and evaluate the outstanding and novel advancements in pattern recognition methods for gas identification. At the same time, this work showcases the recent advancements in utilizing these methods for gas identification, particularly within three crucial domains: ensuring food safety, monitoring the environment, and aiding in medical diagnosis. In conclusion, this study anticipates future research prospects by considering the existing landscape and challenges. It is hoped that this work will make a positive contribution towards mitigating cross-sensitivity in gas-sensitive devices and offer valuable insights for algorithm selection in gas recognition applications. [ABSTRACT FROM AUTHOR]

Dissimilarity measure on intuitionistic fuzzy sets (IFS) are often used in various problems of decision-making, pattern recognition and clustering problems. Numerous dissimilarity measures are constructed based on diverse mathematical concepts. However, not all dissimilarity measure might be suitable for a specific problem or definition. Hence, one should be specific about the nature of the decision to be drawn and choose an appropriate dissimilarity measure. This article aims to classify the dissimilarity measure as moderate, pessimistic and optimistic based on the information held by the IFS. The majority of the currently available dissimilarity measures take into account the moderate viewpoint of information in the IFS. There are a few dissimilarity measures that account for the pessimistic viewpoint of information in the IFSs. However, in some cases, the optimistic viewpoint of the information in the IFSs is also required. In addition, the cross-information factor is an important element of a dissimilarity measure. Consequently, a new dissimilarity measure is being developed that is optimistic and also takes the cross-information factor into account by calculating the change between the maximum and minimum cross-information factors. Furthermore, the dissimilarity measure is designed in such a way that it satisfies the dissimilarity properties under the complete containment relation. Moreover, a method for generating induced dissimilarity measures from a given dissimilarity measure using monotonic functions has been proposed. Numerical and comparative studies demonstrate the out-performance and superiority of the proposed dissimilarity measure, as well as highlights the unique characteristic of an optimistic nature. Finally, the proposed dissimilarity measure is being used to solve variety of application problems, including decision-making, pattern recognition, and clustering. [ABSTRACT FROM AUTHOR]

High accuracy and fast computation time are essential in implementing hand gesture pattern recognition for prosthetic hand using electromyography (EMG) signal. However, several physical parameters affect the characteristics of the EMG signal, including forearm orientation. Therefore, this study aims to develop a deep learning classifier using convolution neural network (CNN) algorithm that maintains accuracy with forearm orientation changes. The main advantage of this method is the simplicity (without feature extraction process) and able to maintain the accuracy against the orientation changes. This method consists of a two-dimensional convolution, max-pooling, four fully connected and output layer. The input layer classifier received six channels of raw EMG signal derived from ten able bodies. As a comparison, several conventional classifiers including support vector machine, k-nearest neighborhood, linear discriminant analysis and decision tree were applied to examine the performance among the classifiers. The result showed that the accuracy of the proposed CNN classifier based on all orientation outperfomed other classifers (96.8 ± 1.87%). Furthermore, the difference in accuracy among the orientations was less then 5%. This indicates that the classifier is able to maintain high accuracy with changes in orientation. In conclusion, this study is applicable in the development of prosthetic hands using EMG signal as control with constant accuracy when the forearm orientation varies. [ABSTRACT FROM AUTHOR]

Obtaining labeled images and samples is a very expensive process and can require intensive labor. At the same time, there are often not enough labeled samples to train an effective classifier. Graph-based semi-supervised methods have attracted much attention in the field because they can use both labeled and unlabeled data. In this letter, we suggest a novel graph-based semi-supervised learning approach that takes full advantage of a small set of labeled samples and a large set of unlabeled samples. We first explain the concept of graph-based semi-supervised learning. Our central proposal is to jointly estimate a low-rank graph together with a latent subspace and soft labels. The proposed system exploits the synergy between the graph information and the latent data representation. This extends and enriches the supervision information in the semi-supervised context and produces a good discriminative linear mapping between the original space and the latent subspace. Several experiments were performed with five image datasets using state-of-the-art methods. The results of the study reveal several noteworthy findings. The proposed framework generally outperforms recent competing approaches in estimating labels and linear embeddings, especially for large datasets such as MNIST. The superiority is maintained even for minimally labeled images, but the recognition accuracy does not always increase. However, in a few cases, the proposed method may be outperformed by other method when the number of labeled samples per class is only one (the dataset UMIST). When we set the number of labeled samples per class to 2, we find that the proposed method consistently performs better than the competing methods. For the extended Yale dataset, the accuracy of the proposed method is 67.08%. This is 1.68% higher than the accuracy of the nearest competing method and an impressive 30.08% higher than the accuracy of the farthest competing method. In the case of the FacePix dataset, the accuracy of the proposed method is 58.08%. This is 2.58% higher than the nearest competing method and 13.58% higher than the farthest competing method. For the UMIST dataset, the accuracy reaches 54.2%, outperforming the nearest competing method by 0.9% and the farthest competing method by 7.5%. We have introduced a novel graph-based semi-supervised method capable of jointly predicting the soft labels and linear embedding and constructing a data graph. The main contribution in this work is the integration of the graph estimation into the objective function of the model. By computing the graph structure while estimating the semi-supervised model, a more optimal solution can be obtained. The synergistic use of data features and adaptive soft labels has indeed contributed to the estimation of a good discriminant model. [ABSTRACT FROM AUTHOR]

Cybersecurity is a critical concern in the transportation sector, where web applications play a pivotal role in managing essential services and sensitive data. Among the various cyber threats, SQL injection attacks pose a significant risk, potentially leading to unauthorized access, data breaches, and disruption of transportation systems. To address this challenge, an advanced approach is proposed that combines Artificial Intelligence (AI) techniques and Natural Language Processing (NLP) to detect and mitigate SQL injection attacks in transportation web applications. In the data collection phase, a comprehensive dataset of real-world attack instances is selected from publicly available sources specializing in cybersecurity datasets. The dataset includes a diverse range of attack vectors and addresses the issue of class imbalance by incorporating both successful and unsuccessful attack attempts. The preprocessing step involves employing NLP techniques to transform the textual input data into a suitable format for AI-based detection. Tokenization, stop-word removal, and stemming are applied to ensure the model effectively analyze and recognize attack patterns. For detection, a logistic regression model is utilized to estimate the probability of a successful SQL injection attack based on the relevant features. Oversampling and undersampling techniques are employed to handle class imbalance and improve the model’s performance. Additionally, feature selection techniques are implemented to reduce noise and enhance pattern recognition. The evaluation of our proposed approach demonstrates a remarkable accuracy detection rate of 99.97%, indicating the model's high capability to identify SQL injection attacks. The precision and recall values further validate the model’s effectiveness in correctly detecting successful attacks and minimizing false positives. The success of our approach lies in its ability to integrate AI and NLP techniques effectively, offering a more robust and reliable solution for detecting and mitigating SQL injection attacks in transportation web applications. By addressing the limitations and exploring future research directions, our approach holds promise in bolstering cybersecurity measures and safeguarding critical transportation infrastructure from evolving cyber threats. [ABSTRACT FROM AUTHOR]

In strategic and radical innovation, the degree of uncertainty and the amount of complexity is much higher compared to 'business as usual'. Therefore, idea management systems are often used to support such innovation processes. An interesting question is what we can learn from studying data in such idea management systems and what potential implications we can derive from the innovation management literature. In this study, we were allowed to access and analyze data from the same idea management system used in 11 radical innovation projects from the years 2012–2018. Our analysis unravels 8 findings that in different ways nuance or challenge current research on innovation management. Finally, we discuss how the integration of artificial intelligence (AI) in idea management systems can support innovation team members in increasing the innovation potential of the ideas that are elaborated. [ABSTRACT FROM AUTHOR]

This work presents a complete conversion scheme for photonic spiking neural networks (SNNs). We verified that the output of an artificial neural network (ANN) trained with the simulated optical activation function can be directly converted into the spike rate of a photonic spiking neuron model. To reveal the feasibility of hardware implementation, we considered the effects of different bit precisions of data and weight, noise level, and bias current mismatch on the converted results. The proposed scheme was evaluated using the Deterding vowel, IRIS, TIDIGITS, and MNIST datasets for pattern recognition, and achieved mean accuracies of 95.80%, 98.67%, 96.19%, and 92.33%, respectively. The proposed scheme can convert an ANN into a photonic SNN with almost no precision loss, and the performance was comparable to that of an ANN trained with the rectified linear unit function. The proposed scheme can enable the high-performance implementation of photonic SNNs. [ABSTRACT FROM AUTHOR]