Narrow Results

The traditional fault detection methods for turntable bearings mainly rely on manual inspection and simple vibration signal analysis. Although these methods can detect faults to a certain extent, they have limitations such as low efficiency, low accuracy, and susceptibility to human factors. To overcome the challenges and limitations of traditional methods, we propose a fault detection method for engineering crane turntable bearings based on the adaptive fireworks algorithm (AFA). Fault detection of turntable bearing of engineering lifting machinery based on an AFA is an innovative method using the fireworks algorithm (FWA) for fault detection. FWA is a kind of optimization algorithm with global search and local search ability, which can effectively solve complex engineering problems. In the fault detection of turntable bearing of engineering lifting machinery, the FWA adaptively adjusts the radius and number of fireworks explosions, so that the algorithm can search in the global scope and detect the fault more accurately. At the same time, the FWA also has a local search ability, which can carry out fine search of the fault area and improve the accuracy of fault detection. By applying the FWA to the fault detection of turntable bearing of engineering lifting machinery, the efficiency and accuracy of fault detection can be effectively improved, the cost of fault detection can be reduced, and the safe operation of engineering lifting machinery can be guaranteed. The fault detection method of turntable bearing of engineering lifting machinery based on an AFA is an innovative method with broad application prospects and can provide an effective solution for the fault detection of engineering lifting machinery.

In the current diversified music creation and consumption environment, the incubation of high-quality music is facing unprecedented challenges, partly due to the significant limitations of traditional beat tracking algorithms in dealing with complex and ever-changing music structures. Therefore, this paper innovatively proposes a real-time music beat tracking algorithm that integrates embedded neural network technology, aiming to break through technical bottlenecks and reshape the paradigm of music feature extraction and classification. The algorithm first conducted in-depth research on the feature extraction and classification technology model of music signals. By integrating embedded neural network technology, deep learning and precise capture of music features have been achieved, effectively overcoming the shortcomings of traditional methods in processing complex music features. On this basis, we further introduced embedded neural networks and utilized their powerful optimization search capabilities to intelligently adjust the data layout for music feature extraction and classification, thereby significantly improving the accuracy of feature extraction and the reliability of classification. To verify the effectiveness of the algorithm, we applied it to real-time detection of rhythm values and precise beat point positions in music. By comparing with international authoritative evaluation datasets such as MIREX2006, the significant improvement of this algorithm in time performance and accuracy was demonstrated.

In order to support precise expression and efficient creation in the animation production process, this paper proposes a feature extraction method for animation script creation by introducing deep learning algorithms from artificial intelligence. First, the basic elements of animation script creation, including screen content, shot motion and time length, were analyzed. Subsequently, in the TF-IDF algorithm, the importance of keywords in the script is quantified by calculating word frequency and inverse text frequency. In the image block sparse representation method, the sparsity degree is used to represent the number of blocks and the target state is described by extracting image features. Finally, using convolutional neural network methods, feature extraction of segmented scripts for animation script creation is achieved through steps such as constructing two-dimensional matrices, performing convolution operations, segment pooling and feature extraction. The experimental results show that the method proposed in this paper has excellent accuracy in extracting features from shot scripts in animation script creation. It can support precise expression and efficient creation in the animation production process, improve the accuracy of feature extraction and provide strong support for the visualization of animation scripts and the design of shot language.

This paper aims to clear up the shortcomings of earlier artwork in extracting cultural factors in free-hand brushwork, particularly in neighborhood contrast processing. We propose an algorithm for extracting cultural elements of free-hand painting based on neighborhood assessment. First, the element and texture of free-hand brushwork are more suitable by nearby histogram equalization, which lays a foundation for subsequent feature extraction. A convolutional neural network (CNN) is then used to robotically analyze and extract key functions from the enhanced photos, along with symbols, shades, texture patterns and identifiers of precise creative styles. These functions are used to efficiently classify free-hand brushwork. The test on massive-scale art database verifies the performance and accuracy of this algorithm inside the function extraction and category of free-hand brushwork. This study now not only presents a brand-new attitude for the evaluation of free-hand brushwork on the technical level, but additionally provides a sensible methodology for the automatic understanding and class of creative works.

The wide range of digital educational resources calls for developing an accurate and efficient method for categorizing and recommending English teaching materials. An automatic classification and recommendation system has been created and implemented using Natural Language Processing (NLP) techniques. Data from essays produced by English Language Learners (ELLs) in grades 8 through 12, as well as components including content, competency levels and score notes, are organized for this study. Models for precise language proficiency assessments are planned to be developed to enhance automated feedback mechanisms. Preprocessing methods such as stop-word removal, stemming and tokenization were applied to tidy up the data. Term Frequency–Inverse Document Frequency (TF–IDF) and word embeddings were two strategies used in the feature extraction process to convert textual data into numerical vectors. Then, the recently created Support Vector Machine–Neural Network–Genetic Algorithm (SVM–NN–GA) was fed to classify these vectors. The model’s performance was evaluated using F1-measure, accuracy, precision and recall metrics. Methods of collaborative filtering and content-based filtering were studied for the recommendation system. In contrast to collaborative filtering, which used user interaction data to identify patterns and suggest relevant items, content-based filtering matched materials with user preferences based on attributes gathered from NLP models. A hybrid recommendation system combines different approaches, increasing recommendations’ personalization and relevance. The results demonstrate that the hybrid recommendation and NLP-based categorization approach as a combination method suggestively improves the effectiveness of selecting appropriate teaching materials, helping teachers to enhance the learning process.

In industrial settings, long-time series data often exhibit high noise levels and complex structures, presenting significant challenges for accurate prediction and maintenance. To address these issues, this paper proposes temporal fusion network (TFN), a novel data fusion algorithm designed for processing industrial time series data. TFN integrates variational mode decomposition (VMD) for denoising, reconstruction, and gap-filling with a hybrid neural network architecture. This architecture combines a temporal convolutional network (TCN) for capturing hierarchical patterns and a gated recurrent unit (GRU) for modeling long-term dependencies. This approach effectively mitigates the influence of high noise and overcomes the limitations of deep convolutional neural network (DCNN) algorithms in handling long-term dependencies. The effectiveness of TFN is demonstrated through experiments on real-world datasets for Industrial Component Degradation Prediction and Predictive Maintenance of Industrial Motors, showcasing its potential for enhancing predictive capabilities in industrial applications.

Speech Emotion Recognition (SER) plays a significant role in human–machine interaction applications. Over the last decade, many SER systems have been anticipated. However, the performance of the SER system remains a challenge owing to the noise, high system complexity and ineffective feature discrimination. SER is challenging and vital, and feature extraction is critical in SER performance. Deep Learning (DL)-based techniques emerge as proficient solutions for SER due to their competence in learning unlabeled data, superior capability of feature representation, capability to handle larger datasets and ability to handle complex features. Different DL techniques, like Long Short-Term Memory (LSTM), Convolutional Neural Network (CNN), Deep Neural Network (DNN) and so on, are successfully presented for automated SER. The study proposes a Robust SER and Classification using the Natural Language Processing with DL (RSERC-NLPDL) model. The presented RSERC-NLPDL technique intends to identify the emotions in the speech signals. In the RSERC-NLPDL technique, pre-processing is initially performed to transform the input speech signal into a valid format. Besides, the RSERC-NLPDL technique extracts a set of features comprising Mel-Frequency Cepstral Coefficients (MFCCs), Zero-Crossing Rate (ZCR), Harmonic-to-Noise Rate (HNR) and Teager Energy Operator (TEO). Next, selecting features can be carried out using Fractal Seagull Optimization Algorithm (FSOA). The Temporal Convolutional Autoencoder (TCAE) model is applied to identify speech emotions, and its hyperparameters are selected using fractal Sand Cat Swarm Optimization (SCSO) algorithm. The simulation analysis of the RSERC-NLPDL method is tested using a speech database. The investigational analysis of the RSERC-NLPDL technique showed superior accuracy outcomes of 94.32% and 95.25% under EMODB and RAVDESS datasets over other models in distinct measures.

Internet of Things (IoT)-assisted consumer electronics refer to common devices that are improved with IoT technology, allowing them to attach to the internet and convey with other devices. These smart devices contain smart home systems, smartphones, wearables, and appliances, which can be monitored remotely, gather, and share data, and deliver advanced functionalities like monitoring, automation, and real-time upgrades. Safety in IoT-assisted consumer electronics signifies a cutting-edge technique to improve device safety and user authentication. Iris recognition (IR) is a biometric authentication technique that employs the exclusive patterns of the iris (the colored part of the eye that surrounds the pupil) to recognize individuals. This method has gained high popularity owing to the uniqueness and stability of iris patterns in finance, healthcare, industries, complex systems, and government applications. With no dual irises being equal and small changes through an individual’s lifetime, IR is considered to be more trustworthy and less susceptible to exterior factors than other biometric detection models. Different classical machine learning (ML)-based IR techniques, the deep learning (DL) approach could not depend on feature engineering and claims outstanding performance. In this paper, we propose an enhanced IR using the Remora fractals optimization algorithm with deep learning (EIR-ROADL) technique for biometric authentication. The main intention of the EIR-ROADL model is to project a hyperparameter-tuned DL technique for automated and accurate IR. For securing consumer electronics, blockchain (BC) technology can be used. In the EIR-ROADL technique, the EIR-ROADL approach uses the Inception v3 method for the feature extraction procedures and its hyperparameter selection process takes place using ROA. For the detection and classification of iris images, the EIR-ROADL technique applies the variational autoencoder (VAE) model. The experimental assessment of the EIR-ROADL algorithm can be executed on benchmark iris datasets. The experimentation outcomes indicated better IR outcomes of the EIR-ROADL methodology with other current approaches and ensured better biometric authentication results.

The categorization and identification of lung disorders in medical imageries are made easier by recent advances in deep learning (DL). As a result, various studies using DL to identify lung illnesses were developed. This study aims to analyze different publications that have been contributed to in order to recognize lung cancer. This literature review examines the many methods for detecting lung cancer. It analyzes several segmentation models that have been used and reviews different research papers. It examines several feature extraction methods, such as those using texture-based and other features. The investigation then concentrates on several cancer detection strategies, including “DL models” and machine learning (ML) models. It is possible to examine and analyze the performance metrics. Finally, research gaps are presented to encourage additional investigation of lung detection models.

Black gram crop belongs to the Fabaceae family and its scientific name is Vigna Mungo.It has high nutritional content, improves the fertility of the soil, and provides atmospheric nitrogen fixation in the soil. The quality of the black gram crop is degraded by diseases such as Yellow mosaic, Anthracnose, Powdery Mildew, and Leaf Crinkle which causes economic loss to farmers and degraded production. The agriculture sector needs to classify plant nutrient deficiencies in order to increase crop quality and yield. In order to handle a variety of difficult challenges, computer vision and deep learning technologies play a crucial role in the agricultural and biological sectors. The typical diagnostic procedure involves a pathologist visiting the site and inspecting each plant. However, manually crop disease assessment is limited due to lesser accuracy and limited access of personnel. To address these problems, it is necessary to develop automated methods that can quickly identify and classify a wide range of plant diseases. In this paper, black gram disease classifications are done through a deep ensemble model with optimal training and the procedure of this technique is as follows: Initially, the input dataset is processed to increase its size via data augmentation. Here, the processes like shifting, rotation, and shearing take place. Then, the model starts with the noise removal of images using median filtering. Subsequent to the preprocessing, segmentation takes place via the proposed deep joint segmentation model to determine the ROI and non-ROI regions. The next process is the extraction of the feature set that includes the features like improved multi-texton-based features, shape-based features, color-based features, and local Gabor X-OR pattern features. The model combines the classifiers like Deep Belief Networks, Recurrent Neural Networks, and Convolutional Neural Networks. For tuning the optimal weights of the model, a new algorithm termed swarm intelligence-based Self-Improved Dwarf Mongoose Optimization algorithm (SIDMO) is introduced. Over the past two decades, nature-based metaheuristic algorithms have gained more popularity because of their ability to solve various global optimization problems with optimal solutions. This training model ensures the enhancement of classification accuracy. The accuracy of the SIDMO, which is around 94.82%, is substantially higher than that of the existing models, which are FPA$=$88.86%, SSOA$=$88.99%, GOA$=$85.84%, SMA$=$85.11%, SRSR$=$85.32%, and DMOA$=$88.99%, respectively.

For different applications, various handcrafted descriptors are reported in the literature. Their results are satisfactory concerning the application they were proposed. Furthermore in the literature, the comparative study discusses these handcrafted descriptors. The main drawback which was noticed in these studies is the restriction of implementation only to single application. This work fills this gap and provides the comparative study of 10 handcrafted for two different applications and these are face recognition (FR) and palmprint recognition (PR). The 10 handcrafted descriptors which are analyzed are local binary pattern (LBP), horizontal elliptical LBP (HELBP), VELBP, robust LBP (RLBP), local phase quantization (LPQ), multiscale block zigzag LBP (MB-ZZLBP), neighborhood mean LBP (NM-LBP), directional threshold LBP (DT-LBP), median robust extended LBP based on neighborhood intensity (MRELBP-NI) and radial difference LBP (RD-LBP). The global feature extraction is performed for all 10 descriptors. PCA and SVMs are used for compaction and matching. Results are done on ORL, GT, IITD-TP and TP. The first two are face datasets and the latter two are palmprint datasets. In face datasets, the descriptor which attains the best recognition accuracy is DT-LBP and in palmprint datasets, it is MB-ZZLBP which surpass the accuracy of the other compared methods.

Sentiment Lexicon (SL) is utilized to extract feedback from a large amount of data, which encompass numerous words, in which a few words contain various semantic definitions in various fields, which is called domain-specific (DS) language, and every word under its domain is demonstrated very significantly as their definition varies from one another. Therefore, this research proposes a Self-configuring knowledge graph-based BiLSTM Classifier with Bat–Harris optimization to determine the accurate sentiment from the comment or text message as well as compute whether it is positive, negative, or neutral, by utilizing the feature extraction process which includes Term Frequency-Inverse Document Frequency (TF-IDF), as well as Hybrid word2vec features, and finally detect the sentiment polarity of text. The experimental results which are depending on the performance metrics show that the developed model is improved compared to the prior method, in which the rate of accuracy of the developed scheme on TP 90 is 93.40% and 94.29%, whereas the sensitivity attains the value of 93.31%, and 94.20%, and finally, the specificity acquires the value of 93.68%, and 94.66% based on datasets 1 and 2.

This study utilizes a combination of machine learning and a global search algorithm to enhance the quality of feature extraction in sports images. A deep residual generative adversarial network is employed to deblur images and sharpen their clarity, while the optimized particle swarm optimization algorithm is utilized to extract image features with precision and identify critical information. According to the experimental results, the research method improves the peak signal-to-noise ratios in ball sports image deblurring by 12.45%, 13.91%, and 17.18%, respectively, and in track and field sports image deblurring by 11.71%, 12.91%, and 21.61%, respectively, when compared with the generative adversarial network, generative adversarial network incorporating the attention mechanism, and multi-scale convolution-based algorithm. The accuracy-recall curve of the particle swarm algorithm that has been optimised for research completely encircles the accuracy-recall curves of the other four algorithms, which verifies the efficacy of the research methodology. The research will offer a more comprehensive perspective on sports image processing.

Phishing is the criminal effort to steal delicate information such as account details, passwords, usernames, credit, and debit card details for malicious use. Phishing fraud might be the most popular cybercrime used today. The online and webmail payment sectors are highly affected by phishing attacks. Nowadays, attackers create many techniques that pave the way for them to steal all personal information from the selected victims easily. However, numerous anti-phishing techniques are used to detect the phishing attack, such as blacklist, visual similarity, heuristic detection, Deep Learning (DL), and Machine Learning (ML) techniques. ML techniques are more efficient at detecting phishing attacks, and these techniques also rectify the drawbacks of existing approaches. This paper provides a detailed review of various phishing techniques, encompassing phishing mediums, phishing vectors, and numerous technical approaches. Also, new research works are analyzed to detect phishing websites using heuristic, visual similarity, DL, and ML models. Neural Networks (NNs), Convolutional Neural Networks (CNNs), Deep Neural Networks (DNNs), Support Vector Machine (SVM), fuzzy logic methods, Long Short-Term Memory (LSTM) techniques, Random Forest (RF), Decision Tree (DT), Adaboost –Extra Tree (AET) classifiers based on ML models are examined in this review paper.

The classification of mangoes’ ripening stage is the major aspect of supplying better fruit grade to buyers, which is a standard necessity of the fruit processing industry. The optical examination, which is manually done, takes to instability, and it involves a lot of work by human workers. To harvest better-quality mangoes, the measurement of maturity is supremely important. During the development and storing at the context temperature, the differentiation in the surface color, size, Total Soluble Solids (TSS) content, firmness and sphericity are analyzed. To tackle the challenges that formed in the classical mango ripening stage, detection approaches are solved by using the newly proposed deep learning approach to identify the maturity state of mangoes. The required mango pictures are taken from the standard databases, and these pictures are given to the image preprocessing to improve image quality and contrast. The improved quality images are applied to the feature extraction section, where the size, shape and color feature are obtained. After that, the ripening of mango is performed through the “Hybrid (1D-2D) Convolution-based Adaptive DenseNet with Attention Mechanism (HCADNet-AM)” to get efficient classification results. The extracted characteristic is applied as the input to the 1D convolution, and the mango images are given as the input for 2D convolution for classifying the maturity stages. The parameter optimization takes place via the Fitness-aided Random Function in Red Panda Optimization (FRF-RPO) during the ripening stage to improve the performance. The research output is validated with conventional ripening techniques to ensure effectiveness.

Fourier transform and entropy are two essential mathematical tools, and they have a fruitful role in system dynamics and machine learning research. In this paper, we propose a generalized composite multiscale amplitude dispersion entropy (GCMADE) and time–frequency dispersion entropy plane. GCMADE measures the complexity of the frequency domain of a time series and can approach the zero complexity of periodic sequences, unlike most other entropy methods. The time–frequency dispersion entropy plane further extracts time domain and frequency domain features of complex signals simultaneously through entropy. Its ability to measure the uncertainty of complex systems is analyzed by simulated data, and the results show that it can effectively distinguish between periodic sequences, chaotic sequences and stochastic processes. Finally, we introduce support vector machine (SVM) to perform mechanical fault diagnosis on five datasets. Compared with the other six algorithms, our method has significantly higher accuracy.

This study introduced a novel auxiliary method for heart failure (HF) diagnosis using the phase space complexity features of ballistocardiogram (BCG) signals collected from piezoelectric sensors. Such a method can potentially monitor high-risk patients out of the clinic. Experimental measurements were collected from 46 patients with HF and 24 healthy subjects. The signals were divided into 1014 nonoverlapping segments (HF: 684 segments, Healthy: 330 segments). First, a digital signal processing framework was established to extract phase space complexity features of BCG with ensemble empirical mode decomposition. Applying a targeted selection strategy, we then identified three key intrinsic mode function (IMF) bands (IMF4–IMF6) for subsequent analysis. Different IMF combinations of features were evaluated using the K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Random Forest (RF), and eXtreme Gradient Boosting (XGB) approaches. Through 10-fold cross-validation, the proposed method exhibited 94.98%, 93.80%, 94.76%, and 94.86% accuracies for the KNN, SVM, RF, and XGB classifiers, respectively. The best performance was achieved by combining IMF4–IMF6 features with the KNN classifier. The proposed BCG signal processing framework is lucrative for diagnosing HF in a home setting.

Recent advancements in structural health monitoring have been significantly driven by the integration of artificial intelligence technologies. This study employs a combination of supervised machine learning techniques, including classification and regression, to accurately detect and localize local thickness reduction defects in a cantilever beams. Our approach utilizes a dataset of 100 signals, comprising 84 defective and 16 healthy states of the beam’s free side displacement, for training machine learning models. Signal processing involves the application of five distinct mode decomposition methods to decompose each signal into its Intrinsic Mode Functions (IMFs). Additionally, four dimensionality reduction methods have been used to reduce the dimensions of the signals. Feature extraction is performed using seven frequency domain, two time domain, and three time–frequency domain methods to capture pertinent patterns and characteristics within the signals. We evaluate the performance of five classification methods and 10 regression methods to predict the location of defects. Our results demonstrate the efficacy of combining specific feature extraction and dimensionality reduction techniques with classification methods, achieving multi-class classification accuracies of up to 99.55%. Moreover, regression methods, particularly the Bayesian ridge regressor, exhibit high accuracy in predicting defect locations, with an $R^2$ value of 99.94% and minimal Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) values. This study highlights the potential of integrating regression and classification-based machine learning approaches for precise damage detection and localization in beam structures.

A convolutional neural network (CNN) based on the Choi–Williams Distribution (CWD) spectrogram and inception module is proposed for radar signal recognition. It is used in a complex electromagnetic environment where the traditional methods have a low recognition rate. The method can automatically identify the modulation type of the received signals while ensuring the multiplicity of extracted features and the precision of signal classification. First, the eight radar signals generated in this research are converted into time–frequency images (TFIs) by CWD. Then the time–frequency information of the signals is extracted by the CNN; two asymmetric four-channel structures are then constructed by Inception structure, and the extracted features are fused separately to achieve the purpose of extracting features in both depth and width of the network. Finally, the classification and recognition of radar signals are realized according to the fused feature map. The experimental results prove that the method has excellent performance in terms of recognition rate and noise immunity compared with the traditional recognition methods.

In this paper, we extend our greedy network-growing algorithm to multi-layered networks. With multi-layered networks, we can solve many complex problems that single-layered networks fail to solve. In addition, the network-growing algorithm is used in conjunction with teacher-directed learning that produces appropriate outputs without computing errors between targets and outputs. Thus, the present algorithm is a very efficient network-growing algorithm. The new algorithm was applied to three problems: the famous vertical-horizontal lines detection problem, a medical data problem and a road classification problem. In all these cases, experimental results confirmed that the method could solve problems that single-layered networks failed to. In addition, information maximization makes it possible to extract salient features in input patterns.