Search

Automatic test paper design is critical in education to reduce workloads for educators and facilitate an efficient teaching process. However, current designs fail to satisfy the realistic teaching requirements of educators, including the consideration of both test quality and efficiency. This is the main reason why teachers still manually construct tests in most teaching environments. In this paper, the quality of tests is quantitatively defined while considering multiple objectives, including a flexible coverage of knowledge points, cognitive levels, and question difficulty. Then, a model based on the technique of linear programming is delicately designed to explore the optimal results for this newly defined problem. However, this technique is not efficient enough, which cannot obtain results in polynomial time. With the consideration of both test quality and generation efficiency, this paper proposes a genetic algorithm (GA) based method, named dynamic programming guided genetic algorithm with adaptive selection (DPGA-AS). In this method, a dynamic programming method is proposed in the population initialization part to improve the efficiency of the genetic algorithm. An adaptive selection method for the GA is designed to avoid prematurely falling into the local optimal for better test quality. The question bank used in our experiments is assembled based on college-level calculus questions from well-known textbooks. The experimental results show that the proposed techniques can construct test papers with both high effectiveness and efficiency. The computation time of the test assembly problem is reduced from 3 hours to 2 seconds for a 5000-size question bank as compared to a linear programming model with similar test quality. The test quality of the proposed method is better than the other baselines.

Aiming at the problem of easy tampering and difficult integrity authentication of paper text documents, this paper proposes a robust content authentication method for printed documents based on text watermarking scheme resisting print-and-scan attack. Firstly, an authentication watermark signal sequence related to content of text document is generated based on the Logistic chaotic map model; then, the authentication watermark signal sequence is embedded into printed paper document by using a robust text watermarking scheme; finally, the watermark information is extracted from scanned image of paper document, and compared with the authentication watermark information calculated in real time by the text document content obtained by OCR technology, thereby performing content integrity authentication of the paper text documents. Experimental results show that our method can achieve the robust content integrity authentication of paper text documents, and can also accurately locate the tampering position. In addition, the document after embedding the watermark information has a good visual effect, and the text watermarking scheme has a large information capacity.

Barreto et al. (2017) showed that the genotypic cyclic competition model with three phenotypes appearing as three morphs of male lizards' throats had the same equilibrium but a wider stability region as the corresponding phenotypic model. In this paper we re-examine stability conditions under the symmetric choice of parameters for the phenotypic model so we can set the same internal equilibrium densities for all three phenotypes. In this setting we compare the stability regions of cyclic allele dominance rule. Next we consider the dynamics on a two-dimensional square lattice space and then show the effect of this spatial structure on the stability of phenotypic model. We obtain the following results: (ⅰ) Cyclic allele dominance rule in a genotypic model gives a wider stable region of internal equilibrium than the allele dominance rule observed in lizards; and (ⅱ) spatial structure drastically changes dynamical behavior, especially when all three phenotypes coexist in almost all the parameter spaces when both competition and dispersal occur locally.

In demanding application scenarios such as clinical psychotherapy and criminal interrogation, the accurate recognition of micro-expressions is of utmost importance but poses significant challenges. One of the main difficulties lies in effectively capturing weak and fleeting facial features and improving recognition performance. To address this fundamental issue, this paper proposed a novel architecture based on a multi-scale 3D residual convolutional neural network. The algorithm leveraged a deep 3D-ResNet50 as the skeleton model and utilized the micro-expression optical flow feature map as the input for the network model. Drawing upon the complex spatial and temporal features inherent in micro-expressions, the network incorporated multi-scale convolutional modules of varying sizes to integrate both global and local information. Furthermore, an attention mechanism feature fusion module was introduced to enhance the model's contextual awareness. Finally, to optimize the model's prediction of the optimal solution, a discriminative network structure with multiple output channels was constructed. The algorithm's performance was evaluated using the public datasets SMIC, SAMM, and CASME Ⅱ. The experimental results demonstrated that the proposed algorithm achieves recognition accuracies of 74.6, 84.77 and 91.35% on these datasets, respectively. This substantial improvement in efficiency compared to existing mainstream methods for extracting micro-expression subtle features effectively enhanced micro-expression recognition performance and increased the accuracy of high-precision micro-expression recognition. Consequently, this paper served as an important reference for researchers working on high-precision micro-expression recognition.

In many studies dealing with mathematical models, the subject is examining the fitting between actual data and the solution of the mathematical model by applying statistical processing. However, if there is a solution that fluctuates greatly due to a small perturbation, it is expected that there will be a large difference between the actual phenomenon and the solution of the mathematical model, even in a short time span. In this study, we address this concern by considering Ulam stability, which is a concept that guarantees that a solution to an unperturbed equation exists near the solution to an equation with bounded perturbations. Although it is known that Ulam stability is guaranteed for the standard von Bertalanffy growth model, it remains unsolved for a model containing the Allee effect. This paper investigates the Ulam stability of a von Bertalanffy growth model with the Allee effect. In a sense, we obtain results that correspond to conditions of the Allee effect being very small or very large. In particular, a more preferable Ulam constant than the existing result for the standard von Bertalanffy growth model, is obtained as the Allee effect approaches zero. In other words, this paper even improves the proof of the result in the absence of the Allee effect. By guaranteeing the Ulam stability of the von Bertalanffy growth model with Allee effect, the stability of the model itself is guaranteed, and, even if a small perturbation is added, it becomes clear that even a small perturbation does not have a large effect on the solutions. Several examples and numerical simulations are presented to illustrate the obtained results.

The cotton-picking robot needs to locate the target object in space in the process of picking in the field and other outdoor strong light complex environments. The difficulty of this process was binocular matching. Therefore, this paper proposes an accurate and fast binocular matching method. This method used the deep learning model to obtain the position and shape of the target object, and then used the matching equation proposed in this paper to match the target object. Matching precision of this method for cotton matching was much higher than that of similar algorithms. It was 54.11, 45.37, 6.15, and 12.21% higher than block matching (BM), semi global block matching (SGBM), pyramid stereo matching network (PSMNet), and geometry and context for deep stereo regression (GC-net) respectively, and its speed was also the fastest. Using this new matching method, the cotton was matched and located in space. Experimental results show the effectiveness and feasibility of the algorithm.

This paper revisits a recently introduced chemostat model of one–species with a periodic input of a single nutrient which is described by a system of delay differential equations. Previous results provided sufficient conditions ensuring the existence and uniqueness of a periodic solution for arbitrarily small delays. This paper partially extends these results by proving—with the construction of Lyapunov–like functions—that the evoked periodic solution is globally asymptotically stable when considering Monod uptake functions and a particular family of nutrient inputs.

Stochastic input-to-state stability (SISS) of the stochastic nonlinear system has received extensive research. This paper aimed to investigate SISS of the stochastic nonlinear system with delayed impulses. First, when all subsystems were stable, using the average impulsive interval method and Lyapunov approach, some theoretical conditions ensuring SISS of the considered system were established. The SISS characteristic of the argumented system with both stable and unstable subsystems was also discussed, then the stochastic nonlinear system with multiple delayed impulse jumps was considered and SISS property was explored. Additionally, it should be noted that the Lyapunov rate coefficient considered in this paper is positively time-varying. Finally, several numerical examples confirmed validity of theoretical results.

Evolutionary multitasking optimization (EMTO) handles multiple tasks simultaneously by transferring and sharing valuable knowledge from other relevant tasks. How to effectively identify transferred knowledge and reduce negative knowledge transfer are two key issues in EMTO. Many existing EMTO algorithms treat the elite solutions in tasks as transferred knowledge between tasks. However, these algorithms may not be effective enough when the global optimums of the tasks are far apart. In this paper, we study an adaptive evolutionary multitasking optimization algorithm based on population distribution information to find valuable transferred knowledge and weaken the negative transfer between tasks. In this paper, we first divide each task population into K sub-populations based on the fitness values of the individuals, and then the maximum mean discrepancy (MMD) is utilized to calculate the distribution difference between each sub-population in the source task and the sub-population where the best solution of the target task is located. Among the sub-populations of the source task, the sub-population with the smallest MMD value is selected, and the individuals in it are used as transferred individuals. In this way, the solution chosen for the transfer may be an elite solution or some other solution. In addition, an improved randomized interaction probability is also included in the proposed algorithm to adjust the intensity of inter-task interactions. The experimental results on two multitasking test suites demonstrate that the proposed algorithm achieves high solution accuracy and fast convergence for most problems, especially for problems with low relevance.

Liver rupture repair surgery serves as one tool to treat liver rupture, especially beneficial for cases of mild liver rupture hemorrhage. Liver rupture can catalyze critical conditions such as hemorrhage and shock. Surgical workflow recognition in liver rupture repair surgery videos presents a significant task aimed at reducing surgical mistakes and enhancing the quality of surgeries conducted by surgeons. A liver rupture repair simulation surgical dataset is proposed in this paper which consists of 45 videos collaboratively completed by nine surgeons. Furthermore, an end-to-end SA-RLNet, a self attention-based recurrent convolutional neural network, is introduced in this paper. The self-attention mechanism is used to automatically identify the importance of input features in various instances and associate the relationships between input features. The accuracy of the surgical phase classification of the SA-RLNet approach is 90.6%. The present study demonstrates that the SA-RLNet approach shows strong generalization capabilities on the dataset. SA-RLNet has proved to be advantageous in capturing subtle variations between surgical phases. The application of surgical workflow recognition has promising feasibility in liver rupture repair surgery.

The Adam algorithm is a common choice for optimizing neural network models. However, its application often brings challenges, such as susceptibility to local optima, overfitting and convergence problems caused by unstable learning rate behavior. In this article, we introduce an enhanced Adam optimization algorithm that integrates Warmup and cosine annealing techniques to alleviate these challenges. By integrating preheating technology into traditional Adam algorithms, we systematically improved the learning rate during the initial training phase, effectively avoiding instability issues. In addition, we adopt a dynamic cosine annealing strategy to adaptively adjust the learning rate, improve local optimization problems and enhance the model's generalization ability. To validate the effectiveness of our proposed method, extensive experiments were conducted on various standard datasets and compared with traditional Adam and other optimization methods. Multiple comparative experiments were conducted using multiple optimization algorithms and the improved algorithm proposed in this paper on multiple datasets. On the MNIST, CIFAR10 and CIFAR100 datasets, the improved algorithm proposed in this paper achieved accuracies of 98.87%, 87.67% and 58.88%, respectively, with significant improvements compared to other algorithms. The experimental results clearly indicate that our joint enhancement of the Adam algorithm has resulted in significant improvements in model convergence speed and generalization performance. These promising results emphasize the potential of our enhanced Adam algorithm in a wide range of deep learning tasks.

A flexible needle has emerged as a crucial clinical technique in contemporary medical practices, particularly for minimally invasive interventions. Its applicability spans diverse surgical domains such as brachytherapy, cardiovascular surgery, neurosurgery and others. Notably, flexible needles find utility in biopsies requiring deep skin penetration to access infected areas. Despite its minimally invasive advantages, the precise guidance of the needle to its intended target, while avoiding damage to bones, blood vessels, organs and tissues, remains a significant challenge for researchers. Consequently, extensive research has been dedicated to enhancing the steering and accuracy of flexible needles. Here, we aim to elucidate the recent advancements, trends and perspectives in flexible needle steering models and path planning over the last 15 years. The discussed models encompass various types, including symmetric-tip needles, curved-tip needles, tendon-actuated needles, programmable needles and the innovative fracture-directed waterjet needles. Moreover, the paper offers a comprehensive analysis, comparing the trajectories followed by these needle models to attain the desired target with minimal tissue damage. By delving into these aspects, the paper contributes to a deeper understanding of the current landscape of flexible needle technology and guides future research directions in this dynamic field.

A stochastic Microcystins degradation model with distributed delay is studied in this paper. We first demonstrate the existence and uniqueness of a global positive solution to the stochastic system. Second, we derive a stochastic critical value $ R_0^s $ related to the basic reproduction number $ R_0 $. By constructing suitable Lyapunov function types, we obtain the existence of an ergodic stationary distribution of the stochastic system if $ R_0^s > 1. $ Next, by means of the method developed to solve the general four-dimensional Fokker-Planck equation, the exact expression of the probability density function of the stochastic model around the quasi-endemic equilibrium is derived, which is the key aim of the present paper. In the analysis of statistical significance, the explicit density function can reflect all dynamical properties of a chemostat model. To validate our theoretical conclusions, we present examples and numerical simulations.

Body posture estimation has been a hot branch in the field of computer vision. This work focuses on one of its typical applications: recognition of various body postures in sports scenes. Existing technical methods were mostly established on the basis of convolution neural network (CNN) structures, due to their strong visual information sensing ability. However, sports scenes are highly dynamic, and many valuable contextual features can be extracted from multimedia frame sequences. To handle the current challenge, this paper proposes a hybrid neural network-based intelligent body posture estimation system for sports scenes. Specifically, a CNN unit and a long short-term memory (LSTM) unit are employed as the backbone network in order to extract key-point information and temporal information from video frames, respectively. Then, a semi-supervised learning-based computing framework is developed to output estimation results. It can make training procedures using limited labeled samples. Finally, through extensive experiments, it is proved that the proposed body posture estimation method in this paper can achieve proper estimation effect in real-world frame samples of sports scenes.

An automated storage and retrieval system (AS/RS) is a key component of enterprise logistics. Its performance metrics include, e.g., order fulfillment time and energy consumption. A crane-based automated storage and retrieval system (CB-AS/RS) is used as the study subject in this paper to build a location allocation model with the goal of minimizing order fulfillment time and minimizing energy consumption. The two-objective problem is transformed into a single-objective problem by the weight method. A genetic algorithm (GA) is used to optimize and simulate the model using spatial mapping coding. A permutation-combination heuristics (PCH) is proposed that follows the coding method and cross-operation of the GA and conducts both arrange-operation and change-operation. During the simulation, the influence of different storage utilization rates and different output and input instruction quantities in a batch of orders on the results is considered. Experimental results show that the results of the PCH algorithm are better than the GA and the optimization results are more stable. In this paper, we provide an optimization idea for the CB-AS/RS researchers and managers.

Bipolar disorder (BD) is a psychiatric disorder that affects an increasing number of people worldwide. The mechanisms of BD are unclear, but some studies have suggested that it may be related to genetic factors with high heritability. Moreover, research has shown that chronic stress can contribute to the development of major illnesses. In this paper, we used bioinformatics methods to analyze the possible mechanisms of chronic stress affecting BD through various aspects. We obtained gene expression data from postmortem brains of BD patients and healthy controls in datasets GSE12649 and GSE53987, and we identified 11 chronic stress-related genes (CSRGs) that were differentially expressed in BD. Then, we screened five biomarkers (IGFBP6, ALOX5AP, MAOA, AIF1 and TRPM3) using machine learning models. We further validated the expression and diagnostic value of the biomarkers in other datasets (GSE5388 and GSE78936) and performed functional enrichment analysis, regulatory network analysis and drug prediction based on the biomarkers. Our bioinformatics analysis revealed that chronic stress can affect the occurrence and development of BD through many aspects, including monoamine oxidase production and decomposition, neuroinflammation, ion permeability, pain perception and others. In this paper, we confirm the importance of studying the genetic influences of chronic stress on BD and other psychiatric disorders and suggested that biomarkers related to chronic stress may be potential diagnostic tools and therapeutic targets for BD.

The utilization of computational models in the field of medical image classification is an ongoing and unstoppable trend, driven by the pursuit of aiding medical professionals in achieving swift and precise diagnoses. Post COVID-19, many researchers are studying better classification and diagnosis of lung diseases particularly, as it was reported that one of the very few diseases greatly affecting human beings was related to lungs. This research study, as presented in the paper, introduces an advanced computer-assisted model that is specifically tailored for the classification of 13 lung diseases using deep learning techniques, with a focus on analyzing chest radiograph images. The work flows from data collection, image quality enhancement, feature extraction to a comparative classification performance analysis. For data collection, an open-source data set consisting of 112,000 chest X-Ray images was used. Since, the quality of the pictures was significant for the work, enhanced image quality is achieved through preprocessing techniques such as Otsu-based binary conversion, contrast limited adaptive histogram equalization-driven noise reduction, and Canny edge detection. Feature extraction incorporates connected regions, histogram of oriented gradients, gray-level co-occurrence matrix and Haar wavelet transformation, complemented by feature selection via regularized neighbourhood component analysis. The paper proposes an optimized hybrid model, improved Aquila optimization convolutional neural networks (CNN), which is a combination of optimized CNN and DENSENET121 with applied batch equalization, which provides novelty for the model compared with other similar works. The comparative evaluation of classification performance among CNN, DENSENET121 and the proposed hybrid model is also done to find the results. The findings highlight the proposed hybrid model's supremacy, boasting 97.00% accuracy, 94.00% precision, 96.00% sensitivity, 96.00% specificity and 95.00% F1-score. In the future, potential avenues encompass exploring explainable machine learning for discerning model decisions and optimizing performance through strategic model restructuring.

In this paper, a fractional-order two delays neural network with ring-hub structure is investigated. Firstly, the stability and the existence of Hopf bifurcation of proposed system are obtained by taking the sum of two delays as the bifurcation parameter. Furthermore, a parameters delay feedback controller is introduced to control successfully Hopf bifurcation. The novelty of this paper is that the characteristic equation corresponding to system has two time delays and the parameters depend on one of them. Selecting two time delays as the bifurcation parameters simultaneously, stability switching curves in $ (\tau_{1}, \tau_{2}) $ plane and crossing direction are obtained. Sufficient criteria for the stability and the existence of Hopf bifurcation of controlled system are given. Ultimately, numerical simulation shows that parameters delay feedback controller can effectively control Hopf bifurcation of system.

As an essential component of mechanical equipment, the fault diagnosis of rolling bearings may not only guarantee the systematic operation of the equipment, but also minimize any financial losses caused by equipment shutdowns. Fault diagnosis algorithms based on convolutional neural networks (CNN) have been widely used. However, traditional CNNs have limited feature representation capabilities, thereby making it challenging to determine their hyperparameters. This paper proposes a fault diagnosis method that combines a 1D-CNN with an attention mechanism and hyperparameter optimization to overcome the aforementioned limitations; this method improves the search speed for optimal hyperparameters of CNN models, improves the diagnostic accuracy, and enhances the representation of fault feature information in CNNs. First, the 1D-CNN is improved by combining it with an attention mechanism to enhance the fault feature information. Second, a swarm intelligence algorithm based on Differential Evolution (DE) and Grey Wolf Optimization (GWO) is proposed, which not only improves the convergence accuracy, but also increases the search efficiency. Finally, the improved 1D-CNN alongside hyperparameters optimization are used to diagnose the faults of rolling bearings. By using the Case Western Reserve University (CWRU) and Jiangnan University (JNU) datasets, when compared to other common diagnosis models, the results demonstrate the usefulness and dependability of the DE-GWO-CNN algorithm in fault diagnosis applications by demonstrating the increased diagnostic accuracy and superior anti-noise capabilities of the proposed method. The fault diagnosis methodology presented in this paper can accurately identify faults and provide dependable fault classification, thereby assisting technicians in promptly resolving faults and minimizing equipment failures and operational instabilities.

Magneto-Acousto-Electrical Tomography (MAET) is a multi-physics coupling imaging modality that integrates the high resolution of ultrasound imaging with the high contrast of electrical impedance imaging. However, the quality of images obtained through this imaging technique can be easily compromised by environmental or experimental noise, thereby affecting the overall quality of the imaging results. Existing methods for magneto-acousto-electrical image denoising lack the capability to model local and global features of magneto-acousto-electrical images and are unable to extract the most relevant multi-scale contextual information to model the joint distribution of clean images and noise images. To address this issue, we propose a Dual Generative Adversarial Network based on Attention Residual U-Net (ARU-DGAN) for magneto-acousto-electrical image denoising. Specifically, our model approximates the joint distribution of magneto-acousto-electrical clean and noisy images from two perspectives: noise removal and noise generation. First, it transforms noisy images into clean ones through a denoiser; second, it converts clean images into noisy ones via a generator. Simultaneously, we design an Attention Residual U-Net (ARU) to serve as the backbone of the denoiser and generator in the Dual Generative Adversarial Network (DGAN). The ARU network adopts a residual mechanism and introduces a linear Self-Attention based on Cross-Normalization (CNorm-SA), which is proposed in this paper. This design allows the model to effectively extract the most relevant multi-scale contextual information while maintaining high resolution, thereby better modeling the local and global features of magneto-acousto-electrical images. Finally, extensive experiments on a real-world magneto-acousto-electrical image dataset constructed in this paper demonstrate significant improvements in preserving image details achieved by ARU-DGAN. Furthermore, compared to the state-of-the-art competitive methods, it exhibits a 0.3 dB increase in PSNR and an improvement of 0.47% in SSIM.

Fire incidents near power transmission lines pose significant safety hazards to the regular operation of the power system. Therefore, achieving fast and accurate smoke detection around power transmission lines is crucial. Due to the complexity and variability of smoke scenarios, existing smoke detection models suffer from low detection accuracy and slow detection speed. This paper proposes an improved model for smoke detection in high-voltage power transmission lines based on the improved YOLOv7-tiny. First, we construct a dataset for smoke detection in high-voltage power transmission lines. Due to the limited number of real samples, we employ a particle system to randomly generate smoke and composite it into randomly selected real scenes, effectively expanding the dataset with high quality. Next, we introduce multiple parameter-free attention modules into the YOLOv7-tiny model and replace regular convolutions in the Neck of the model with Spd-Conv (Space-to-depth Conv) to improve detection accuracy and speed. Finally, we utilize the synthesized smoke dataset as the source domain for model transfer learning. We pre-train the improved model and fine-tune it on a dataset consisting of real scenarios. Experimental results demonstrate that the proposed improved YOLOv7-tiny model achieves a 2.61% increase in mean Average Precision (mAP) for smoke detection on power transmission lines compared to the original model. The precision is improved by 2.26%, and the recall is improved by 7.25%. Compared to other object detection models, the smoke detection proposed in this paper achieves high detection accuracy and speed. Our model also improved detection accuracy on the already publicly available wildfire smoke dataset Figlib (Fire Ignition Library).

Aiming at the personal credit evaluation of commercial banks, this paper constructs a classified prediction model based on machine learning methods to predict the default risk. At the same time, this paper proposes to combine the sparrow search algorithm (SSA) with the support vector machine (SVM) to explore the application of the SSA-SVM model in personal default risk prediction. Therefore, this paper takes the personal credit data as the original data, carries out statistical analysis, normalization and principal factor analysis, and substitutes the obtained variables as independent variables into the SSA-SVM model. Under the premise of the same model, the experimental results show that the evaluation indexes of the experimental data are better than the original data, which shows that it is effective for the data processing operation of the original data in this paper. On the premise of the same data, each evaluation index of the SSA-SVM model is better than the SVM model, which shows that the hybridized model established in this paper is better than the latter one in predicting personal default risk, and has certain practical value.

Shortages of medical equipment, growth in medical waste and carbon emissions have increased healthcare pressures and has a huge impact on the environment. An efficient scheduling of medical equipment will effectively reduce the pressure on healthcare and improve the healthcare system's ability to respond to unexpected disasters. A medical equipment scheduling system was established to improve the sustainable utilization of medical equipment within the healthcare network and to reduce the carbon emissions of the healthcare process. First, this paper combines medical equipment information to establish a medical equipment scheduling decision model that considers pollution to filter qualified medical equipment for scheduling. Then, this paper constructs and solves a multi-objective robust optimization model by collecting the patient's travel information and the medical pressure information of each region. In addition, to meet dynamic healthcare needs, a dynamic medical equipment configuration framework was constructed to enhance the flexibility of equipment scheduling and the resilience of the healthcare network. Combined with case studies, the results show that the medical equipment scheduling system can help decision makers make quick scheduling decisions and achieve sustainable use of medical equipment, with a corresponding increase in medical equipment utilization of 12.25% and a reduction in carbon emissions of 26.50%. The study will help enhance healthcare resource utilization and contribute to the net-zero goal of green healthcare.

In order to obtain an optimal trajectory for indoor AGV, this paper combined an improved ACO and fuzzy DWA (IACO-DWA) algorithm, which can provide an optimal path with collision-free under higher optimization efficiency. The highlights of this paper are detailed as follows: Firstly, an improved adaptive pseudo-random transition strategy is adopted in the state transition probability with an angle factor. A reward and punishment mechanism is introduced in the pheromone updating strategy, then a path optimization strategy called IACO is proposed for the more optimized path. Secondly, IDWA adopted three fuzzy controllers of direction, security and adjustment coefficients through evaluating directional and safety principles, then improving the angular velocity by processing the linear velocity with linear normalization. By adapting to the changes of the environment, the IDWA parameters can be dynamically adjusted to ensure the optimal running speed and reasonable path of AGV. Thirdly, aiming to deal with the path-planning problem in complex environments, we combined IACO with IDWA, the hybrid algorithm involves dividing the globally optimal path obtained from IACO planning into multiple virtual sub-target points. IDWA completes the path planning by switching between these local target points, thereby improving the efficiency of the path planning. Finally, simulations is verified by Matlab and experiment results on the QBot2e platform are given to verify IACO-DWA algorithm's effectiveness and high performance.

Prescription data is an important focus and breakthrough in the study of clinical treatment rules, and the complex multidimensional relationships between Traditional Chinese medicine (TCM) prescription data increase the difficulty of extracting knowledge from clinical data. This paper proposes a complex prescription recognition algorithm (MTCMC) based on the classification and matching of TCM prescriptions with classical prescriptions to identify the classical prescriptions contained in the prescriptions and provide a reference for mining TCM knowledge. The MTCMC algorithm first calculates the importance level of each drug in the complex prescriptions and determines the core prescription combinations of patients through the Analytic Hierarchy Process (AHP) combined with drug dosage. Secondly, a drug attribute tagging strategy was used to quantify the functional features of each drug in the core prescriptions; finally, a Bidirectional Long Short-Term Memory Network (BiLSTM) was used to extract the relational features of the core prescriptions, and a vector representation similarity matrix was constructed in combination with the Siamese network framework to calculate the similarity between the core prescriptions and the classical prescriptions. The experimental results show that the accuracy and F1 score of the prescription matching dataset constructed based on this paper reach 94.45% and 94.34% respectively, which is a significant improvement compared with the models of existing methods.

Health literacy refers to the ability of individuals to obtain and understand health information and use it to maintain and promote their own health. This paper manages to make predictions toward its development degree in society with use of a big data-driven statistical learning method. Actually, such results can be analyzed by discovering latent rules from massive public textual contents. As a result, this paper proposes a deep information fusion-based smart prediction approach for health literacy. Specifically, the latent Dirichlet allocation (LDA) and convolutional neural network (CNN) structures are utilized as the basic backbone to understand semantic features of textual contents. The feature learning results of LDA and CNN can be then mapped into prediction results via following multi-dimension computing structures. After constructing the CNN model, we can input health information into the model for feature extraction. The CNN model can automatically learn valuable features from raw health information through multi-layer convolution and pooling operations. These characteristics may include lifestyle habits, physiological indicators, biochemical indicators, etc., reflecting the patient's health status and disease risk. After extracting features, we can train the CNN model through a training set and evaluate the performance of the model using a test set. The goal of this step is to optimize the parameters of the model so that it can accurately predict health information. We can use common evaluation indicators such as accuracy, precision, recall, etc. to evaluate the performance of the model. At last, some simulation experiments are conducted on real-world data collected from famous international universities. The case study analyzes health literacy difference between China of developed countries. Some prediction results can be obtained from the case study. The proposed approach can be proved effective from the discussion of prediction results.

After over three years of COVID-19, it has become clear that infectious diseases are difficult to eradicate, and humans remain vulnerable under their influence in a long period. The presence of presymptomatic and asymptomatic patients is a significant obstacle to preventing and eliminating infectious diseases. However, the long-term transmission of infectious diseases involving asymptomatic patients still remains unclear. To address this issue, this paper develops a novel Markov process for infectious diseases with asymptomatic patients by means of a continuous-time level-dependent quasi-birth-and-death (QBD) process. The model accurately captures the transmission of infectious diseases by specifying several key parameters (or factors). To analyze the role of asymptomatic and symptomatic patients in the infectious disease transmission process, a simple sufficient condition for the stability of the Markov process of infectious diseases is derived using the mean drift technique. Then, the stationary probability vector of the QBD process is obtained by using RG-factorizations. A method of using the stationary probability vector is provided to obtain important performance measures of the model. Finally, some numerical experiments are presented to demonstrate the model's feasibility through analyzing COVID-19 as an example. The impact of key parameters on the system performance evaluation and the infectious disease transmission process are analyzed. The methodology and results of this paper can provide theoretical and technical support for the scientific control of the long-term transmission of infectious diseases, and we believe that they can serve as a foundation for developing more general models of infectious disease transmission.

With the rapid development of e-commerce, express delivery has been chosen and accepted by consumers, and a large number of express packages have resulted in serious waste of resources and environmental pollution. Because of the irregularity of online goods purchases by users in real life, logistics parks are unable to accurately judge the recycling needs of various regions. In order to solve this problem, we propose an improved empirical mode decomposition (IEMD) algorithm combined with a long-short-term memory (LSTM) network to deal with the addresses and categories in logistics data, analyze the distribution of recyclable logistics waste in the logistics park service area and in the express recycling station within the logistics park, judge the value of recyclable logistics waste, optimize the best path for recycling vehicles and improve the success rate of logistics waste recycling. In order to better research and verify the IEMD-LSTM prediction model, we model and simulate the algorithm behavior of the express waste packaging recycling prediction model system, and compare it with other classification methods through specific logistics data experiments. The prediction accuracy, stability and advantages of the four algorithms are analyzed and compared, and the application reliability of the algorithm proposed in this paper to the logistics waste recycling process is verified. The application in the actual express logistics packaging recycling case shows the feasibility and effectiveness of the waste recycling scheme proposed in this paper.

With the help of advanced technology, the automotive industry is in continuous evolution. Modern vehicles are not only comprised of mechanical components but also contain highly complex electronic devices and connections to the outside world. Today's vehicle usually has between 30 and 70 ECUs (Electronic Control Units), which communicate with each other over standard communication protocols. There are different types of in-vehicle network protocols and bus systems, including the Controlled Area Network (CAN), Local Interconnected Network (LIN), FlexRay, Media Oriented System Transport (MOST), and Automotive Ethernet (AE). Modern cars are also able to communicate with other devices through wired or wireless interfaces such as USB, Bluetooth, Wi-Fi or even 5G. Such interfaces may expose the internal network to the outside world and can be seen as entry points for cyber-attacks. In this paper, the main interest is in the AE network protocol. AE is a special Ethernet design that provides the bandwidth needed for today's applications, and the potential for even greater performance in the future. However, AE is a "best effort" protocol, which cannot be considered reliable. This implies that it is not trustworthy in terms of reliability and timely deliveries. The focus of this paper is to present a state-of-the-art survey of security threats and protection mechanisms relating to AE. After introducing and comparing the different protocols being used in the embedded networks of current vehicles, we analyze the potential threats targeting the AE network and describe how attackers' opportunities can be enhanced by the new communication abilities of modern cars. Finally, we present and compare the AE security solutions currently being devised to address these problems and propose some recommendations and challenges to deal with security issue in AE protocol.

The purpose of infrared and visible image fusion is to integrate the complementary information from heterogeneous images in order to enhance their detailed scene information. However, existing deep learning fusion methods suffer from an imbalance between fusion performance and computational resource consumption. Additionally, fusion layers or fusion rules fail to effectively combine heteromodal feature information. To address these challenges, this paper presents a novel algorithm called infrared and visible image fusion network base on fast edge convolution (FECFusion). During the training phase, the proposed algorithm enhances the extraction of texture features in the source image through the utilization of structural re-parameterization edge convolution (RECB) with embedded edge operators. Subsequently, the attention fusion module (AFM) is employed to sufficiently fuze both unique and public information from the heteromodal features. In the inference stage, we further optimize the training network using the structural reparameterization technique, resulting in a VGG-like network architecture. This optimization improves the fusion speed while maintaining the fusion performance. To evaluate the performance of the proposed FECFusion algorithm, qualitative and quantitative experiments are conducted. Seven advanced fusion algorithms are compared using MSRS, TNO, and M3FD datasets. The results demonstrate that the fusion algorithm presented in this paper achieves superior performance in multiple evaluation metrics, while consuming fewer computational resources. Consequently, the proposed algorithm yields better visual results and provides richer scene detail information.

In the intelligent manufacturing environment, modern industry is developing at a faster pace, and there is an urgent need for reasonable production scheduling to ensure an organized production order and a dependable production guarantee for enterprises. Additionally, production cooperation between enterprises and different branches of enterprises is increasingly common, and distributed manufacturing has become a prevalent production model. In light of these developments, this paper presents the research background and current state of distributed shop scheduling. It summarizes relevant research on issues that align with the new manufacturing model, explores hot topics and concerns and focuses on the classification of distributed parallel machine scheduling, distributed flow shop scheduling, distributed job shop scheduling and distributed assembly shop scheduling. The paper investigates these scheduling problems in terms of single-objective and multi-objective optimization, as well as processing constraints. It also summarizes the relevant optimization algorithms and their limitations. It also provides an overview of research methods and objects, highlighting the development of solution methods and research trends for new problems. Finally, the paper analyzes future research directions in this field.

In this paper, we study a predator-prey system, the modified Holling-Tanner model with strong Allee effect. The existence and stability of the non-negative equilibria are discussed first. Several kinds of bifurcation phenomena, which the model may undergo, such as saddle-node bifurcation, Hopf bifurcation, and Bogdanov-Takens bifurcation, are studied second. Bifurcation diagram for Bogdanov-Takens bifurcation of codimension 2 is given. Then, possible dynamical behaviors of this model are illustrated by numerical simulations. This paper appears to be the first study of the modified Holling-Tanner model that includes the influence of a strong Allee effect.

The ongoing emergence of COVID-19 and the maturation of cold chain technology, have aided in the rapid development of the fresh produce e-commerce industry. Taking into account the characteristics of consumers' demand for fresh products, this paper constructs a location allocation model of a front warehouse for fresh e-commerce with the objective of minimizing the total cost. An improved immune optimization algorithm is proposed in this paper, and the effectiveness of the proposed algorithm is demonstrated by a real case study. The results show that the improved immune optimization algorithm outperforms the traditional genetic algorithm in terms of solution accuracy; the proposed location model can effectively help fresh produce e-commerce enterprises open new front-end warehouses when demand is increasing, as well as provide optimal economic decision-making for front warehouse layout.

With the rapid development of smart campus, this paper studies the attitude tracking control of flexible manipulator (FM) in colleges and universities under elastic vibration and external disturbances. First, different from the traditional modeling based on ordinary differential equations (ODEs), the partial differential equations (PDEs) dynamic model of a manipulator system is established based on the Hamilton principle (HP). Second, the boundary control condition of the end system of the manipulator is introduced to adjust the vibration of the manipulator. Furthermore, a Proportional-Derivative (PD) boundary control (PDBC) strategy is proposed by the Lyapunov function to suppress the vibration of the manipulator. Finally, a numerical comparison simulation based on MATLAB/SIMULINK further verifies the robustness and anti-disturbance performance of the control method proposed in this paper.

Due to the coupling effect of external environmental noise and vibration noise, the feature rate of the original hydroelectric unit fault signal is not prominent, which will affect the performance of fault diagnosis algorithms. To solve the above problems, this paper proposes a PSO-MCKD-MFFResnet algorithm for fault diagnosis of hydropower units (Particle swarm optimization, PSO; Maximum correlation kurtosis deconvolution, MCKD; Multi-scale feature fusion residual network, MFFResnet). In practical applications, the selection of key parameters in the traditional MCKD method is heavily dependent on prior knowledge. First, this paper proposes a PSO-MCKD enhancement algorithm for fault features, which uses the PSO algorithm to search for the influencing parameters of MCKD to enhance the features from the original fault signal. Second, a fault feature diagnosis algorithm based on MFFResnet is proposed to improve the utilization of local features. The multi-scale residual module is used to extract features at different scales and then put the enhanced signal into MFFResnet for training and classification. The experimental results show that our approach can accurately and effectively classify the fault types of hydropower units, with an accuracy rate of 98.85. It is superior to other representative algorithms in different indicators and has a good stability.

The utilization of intelligent computing in digital teaching quality evaluation has been a practical demand in smart cities. Currently, related research works can be categorized into two types: textual data-based approaches and visual data-based approaches. Due to the gap between their different formats and modalities, it remains very challenging to integrate them together when conducting digital teaching quality evaluation. In fact, the two types of information can both reflect distinguished knowledge from their own perspectives. To bridge this gap, this paper proposes a textual and visual features-jointly driven hybrid intelligent system for digital teaching quality evaluation. Visual features are extracted with the use of a multiscale convolution neural network by introducing receptive fields with different sizes. Textual features serve as the auxiliary contents for major visual features, and are extracted using a recurrent neural network. At last, we implement the proposed method through some simulation experiments to evaluate its practical running performance, and a real-world dataset collected from teaching activities is employed for this purpose. We obtain some groups of experimental results, which reveal that the hybrid intelligent system developed by this paper can bring more than 10% improvement of efficiency towards digital teaching quality evaluation.

Verification is the only way to make sure if a node is influenced or not because of the uncertainty of information diffusion in the temporal contact network. In the previous methods, only $ N $ influenced nodes could be found for a given number of verifications $ N $. The target of discovering influenced nodes is to find more influenced nodes with the limited number of verifications. To tackle this difficult task, the common nodes on the temporal diffusion paths is proposed in this paper. We prove that if a node $ v $ is confirmed as the influenced node and there exist common nodes on the temporal diffusion paths from the initial node to the node $ v $, these common nodes can be regarded as the influenced nodes without verification. It means that it is possible to find more than $ N $ influenced nodes given $ N $ verifications. The common nodes idea is applied to search influenced nodes in the temporal contact network, and three algorithms are designed based on the idea in this paper. The experiments show that our algorithms can find more influenced nodes in the existence of common nodes.

In 3D reconstruction tasks, camera parameter matrix estimation is usually used to present the single view of an object, which is not necessary when mapping the 3D point to 2D image. The single view reconstruction task should care more about the quality of reconstruction instead of the alignment. So in this paper, we propose an implicit field knowledge distillation model (IFKD) to reconstruct 3D objects from the single view. Transformations are performed on 3D points instead of the camera and keep the camera coordinate identified with the world coordinate, so that the extrinsic matrix can be omitted. Besides, a knowledge distillation structure from 3D voxel to the feature vector is established to further refine the feature description of 3D objects. Thus, the details of a 3D model can be better captured by the proposed model. This paper adopts ShapeNet Core dataset to verify the effectiveness of the IFKD model. Experiments show that IFKD has strong advantages in IOU and other core indicators compared with the camera matrix estimation methods, which verifies the feasibility of the new proposed mapping method.

This paper considers the distributed tracking control problem for a class of nonlinear multi-agent systems with nonlinearly parameterized control coefficients and inherent nonlinearities. The essential of multi-agent systems makes it difficult to directly generalize the existing works for single nonlinearly parameterized systems with uncontrollable unstable linearization to the case in this paper. To dominate the inherent nonlinearities and nonlinear parametric uncertainties, a powerful distributed adaptive tracking control is presented by combing the algebra graph theory with the distributed backstepping method, which guarantees that all the closed-loop system signals are global bounded while the range of the tracking error between the follower's output and the leader's output can be tuned arbitrarily small. Finally, a numerical example is provided to verify the validity of the developed methods.

In this paper, considering the aggregation effect and Allee effect of cyanobacteria populations and the harvesting of both cyanobacteria and fish by human beings, a new cyanobacteria-fish model with two harvesting terms and a modified Holling type Ⅳ functional response function is proposed. The main purpose of this paper is to further elucidate the influence of harvesting terms on the dynamic behavior of a cyanobacteria-fish model. Critical conditions for the existence and stability of several interior equilibria are given. The economic equilibria and the maximum sustainable total yield problem are also studied. The model exhibits several bifurcations, such as transcritical bifurcation, saddle-node bifurcation, Hopf bifurcation and Bogdanov-Takens bifurcation. It is concluded from a biological perspective that the survival mode of cyanobacteria and fish can be determined by the harvesting terms. Finally, concrete examples of our model are given through numerical simulations to verify and enrich the theoretical results.

In this paper, the exponential bipartite consensus issue is investigated for multi-agent networks, whose dynamic is characterized by fractional diffusion partial differential equations (PDEs). The main contribution is that a novel exponential convergence principle is proposed for networks of fractional PDEs via aperiodically intermittent control scheme. First, under the aperiodically intermittent control strategy, an exponential convergence principle is developed for continuously differentiable function. Second, on the basis of the proposed convergence principle and the designed intermittent boundary control protocol, the exponential bipartite consensus condition is addressed in the form of linear matrix inequalities (LMIs). Compared with the existing works, the result of the exponential intermittent consensus presented in this paper is applied to the networks of PDEs. Finally, the high-speed aerospace vehicle model is applied to verify the effectiveness of the control protocol.

This paper investigates the fault tolerance problem of flexible satellites subject to actuator faults and multiple disturbances. An adaptive sliding mode fault tolerant control (ASMFTC) approach based on Takagi-Sugeno (T-S) fuzzy disturbance observer (TSFDO) is presented for attitude control system (ACS) under loss of actuator effectiveness, environmental disturbance torque and elastic modal generated by flexible appendages. Compared with the traditional disturbance observer based control (DOBC) methods, the T-S fuzzy technology is applied to estimate the unknown nonlinear elastic modal. Then, the energy bounded disturbance is eliminated by designing an adaptive sliding mode controller. The proposed ASMFTC design can guarantee the sliding surface to approach zero. Finally, the effectiveness of the control method proposed in this paper is further verified by comparative simulation.

There are approximately 2.2 billion people around the world with varying degrees of visual impairments. Among them, individuals with severe visual impairments predominantly rely on hearing and touch to gather external information. At present, there are limited reading materials for the visually impaired, mostly in the form of audio or text, which cannot satisfy the needs for the visually impaired to comprehend graphical content. Although many scholars have devoted their efforts to investigating methods for converting visual images into tactile graphics, tactile graphic translation fails to meet the reading needs of visually impaired individuals due to image type diversity and limitations in image recognition technology. The primary goal of this paper is to enable the visually impaired to gain a greater understanding of the natural sciences by transforming images of mathematical functions into an electronic format for the production of tactile graphics. In an effort to enhance the accuracy and efficiency of graph element recognition and segmentation of function graphs, this paper proposes an MA Mask R-CNN model which utilizes MA ConvNeXt as its improved feature extraction backbone network and MA BiFPN as its improved feature fusion network. The MA ConvNeXt is a novel feature extraction network proposed in this paper, while the MA BiFPN is a novel feature fusion network introduced in this paper. This model combines the information of local relations, global relations and different channels to form an attention mechanism that is able to establish multiple connections, thus increasing the detection capability of the original Mask R-CNN model on slender and multi-type targets by combining a variety of multi-scale features. Finally, the experimental results show that MA Mask R-CNN attains an 89.6% mAP value for target detection and 72.3% mAP value for target segmentation in the instance segmentation of function graphs. This results in a 9% mAP improvement for target detection and 12.8% mAP improvement for target segmentation compared to the original Mask R-CNN.

In order to further promote the application and development of unmanned aviation in the manned field, and reduce the difficulty that airlines cannot avoid due to unexpected factors such as bad weather, aircraft failure, and so on, the problem of restoring aircraft routes has been studied. To reduce the economic losses caused by flight interruption, this paper divides the repair problem of aircraft operation plans into two sub problems, namely, the generation of flight routes and the reallocation of aircraft. Firstly, the existing fixed-point iteration method proposed by Dang is used to solve the feasible route generation model based on integer programming. To calculate quickly and efficiently, a segmentation method that divides the solution space into mutually independent segments is proposed as the premise of distributed computing. The feasible route is then allocated to the available aircraft to repair the flight plan. The experimental results of two examples of aircraft fault grounding and airport closure show that the method proposed in this paper is effective for aircraft route restoration.

Sparse mobile crowd sensing saves perception cost by recruiting a small number of users to perceive data from a small number of sub-regions, and then inferring data from the remaining sub-regions. The data collected by different people on their respective trajectories have different values, and we can select participants who can collect high-value data based on their trajectory predictions. In this paper, we study two aspects of user trajectory prediction and user recruitment. First, we propose an STGCN-GRU user trajectory prediction algorithm, which uses the STGCN algorithm to extract features related to temporal and spatial information from the trajectory map, and then inputs the feature sequences into GRU for trajectory prediction, and this algorithm improves the accuracy of user trajectory prediction. Second, an ADQN (action DQN) user recruitment algorithm is proposed.The ADQN algorithm improves the objective function in DQN on the idea of reinforcement learning. The action with the maximum input value is found from the Q network, and then the output value of the objective function of the corresponding action Q network is found. This reduces the overestimation problem that occurs in Q networks and improves the accuracy of user recruitment. The experimental results show that the evaluation metrics FDE and ADE of the STGCN-GRU algorithm proposed in this paper are better than other representative algorithms. And the experiments on two real datasets verify the effectiveness of the ADQN user selection algorithm, which can effectively improve the accuracy of data inference under budget constraints.

The objective of this paper is to design a fast and efficient immune algorithm for solving various optimization problems. The immune algorithm (IA), which simulates the principle of the biological immune system, is one of the nature-inspired algorithms and its many advantages have been revealed. Although IA has shown its superiority over the traditional algorithms in many fields, it still suffers from the drawbacks of slow convergence and local minima trapping problems due to its inherent stochastic search property. Many efforts have been done to improve the search performance of immune algorithms, such as adaptive parameter setting and population diversity maintenance. In this paper, an improved immune algorithm (IIA) which utilizes a parallel mutation mechanism (PM) is proposed to solve the Lennard-Jones potential problem (LJPP). In IIA, three distinct mutation operators involving cauchy mutation (CM), gaussian mutation (GM) and lateral mutation (LM) are conditionally selected to be implemented. It is expected that IIA can effectively balance the exploration and exploitation of the search and thus speed up the convergence. To illustrate its validity, IIA is tested on a two-dimension function and some benchmark functions. Then IIA is applied to solve the LJPP to exhibit its applicability to the real-world problems. Experimental results demonstrate the effectiveness of IIA in terms of the convergence speed and the solution quality.

The recognition of traffic signs is of great significance to intelligent driving and traffic systems. Most current traffic sign recognition algorithms do not consider the impact of rainy weather. The rain marks will obscure the recognition target in the image, which will lead to the performance degradation of the algorithm, a problem that has yet to be solved. In order to improve the accuracy of traffic sign recognition in rainy weather, we propose a rainy traffic sign recognition algorithm. The algorithm in this paper includes two modules. First, we propose an image deraining algorithm based on the Progressive multi-scale residual network (PMRNet), which uses a multi-scale residual structure to extract features of different scales, so as to improve the utilization rate of the algorithm for information, combined with the Convolutional long-short term memory (ConvLSTM) network to enhance the algorithm's ability to extract rain mark features. Second, we use the CoT-YOLOv5 algorithm to recognize traffic signs on the recovered images. In this paper, in order to improve the performance of YOLOv5 (You-Only-Look-Once, YOLO), the 3 × 3 convolution in the feature extraction module is replaced by the Contextual Transformer (CoT) module to make up for the lack of global modeling capability of Convolutional Neural Network (CNN), thus improving the recognition accuracy. The experimental results show that the deraining algorithm based on PMRNet can effectively remove rain marks, and the evaluation indicators Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM) are better than the other representative algorithms. The mean Average Precision (mAP) of the CoT-YOLOv5 algorithm on the TT100k datasets reaches 92.1%, which is 5% higher than the original YOLOv5.

Environmental wireless sensor networks (EWSNs) are essential in environmental monitoring and are widely used in gas monitoring, soil monitoring, natural disaster early warning and other fields. EWSNs are limited by the sensor battery capacity and data collection range, and the usual deployment method is to deploy many sensor nodes in the monitoring zone. This deployment method improves the robustness of EWSNs, but introduces many redundant nodes, resulting in a problem of duty cycle design, which can be effectively solved by duty cycle optimization. However, the duty cycle optimization in EWSNs is an NP-Hard problem, and the complexity of the problem increases exponentially with the number of sensor nodes. In this way, non-heuristic algorithms often fail to obtain a deployment solution that meets the requirements in reasonable time. Therefore, this paper proposes a novel heuristic algorithm, the Quantum Evolutionary Golden Jackal Optimization Algorithm (QEGJOA), to solve the duty cycle optimization problem. Specifically, QEGJOA can effectively prolong the lifetime of EWSNs by duty cycle optimization and can quickly get a deployment solution in the face of multi-sensor nodes. New quantum exploration and exploitation operators are designed, which greatly improves the global search ability of the algorithm and enables the algorithm to effectively solve the problem of excessive complexity in duty cycle optimization. In addition, this paper designs a new sensor duty cycle model, which has the advantages of high accuracy and low complexity. The simulation shows that the QEGJOA proposed in this paper improves by 18.69, 20.15 and 26.55 compared to the Golden Jackal Optimization (GJO), Whale Optimization Algorithm (WOA) and the Simulated Annealing Algorithm (SA).

The semiconductor manufacturing industry relies heavily on wafer surface defect detection for yield enhancement. Machine learning and digital image processing technologies have been used in the development of various detection algorithms. However, most wafer surface inspection algorithms are not be applied in industrial environments due to the difficulty in obtaining training samples, high computational requirements, and poor generalization. In order to overcome these difficulties, this paper introduces a full-flow inspection method based on machine vision to detect wafer surface defects. Starting with the die image segmentation stage, where a die segmentation algorithm based on candidate frame fitting and coordinate interpolation is proposed for die sample missing matching segmentation. The method can segment all the dies in the wafer, avoiding the problem of missing dies splitting. After that, in the defect detection stage, we propose a die defect anomaly detection method based on defect feature clustering by region, which can reduce the impact of noise in other regions when extracting defect features in a single region. The experiments show that the proposed inspection method can precisely position and segment die images, and find defective dies with an accuracy of more than 97%. The defect detection method proposed in this paper can be applied to inspect wafer manufacturing.