This PDF file contains the front matter associated with SPIE Proceedings Volume 12290, including the Title Page, Copyright information, Table of Contents, and Committee Pages.

Based on achievement efficiency and time efficiency, this paper establishes the atomic attack efficiency model, and constructs the network attack efficiency model and the network attack chain efficiency model. In this model, to build a network host information security integrity test tree, leaf nodes can be considered as the vulnerabilities of each host in the target network, edges between the nodes can be seen as the attack methods of vulnerabilities, middle nodes can be considered as the attack results of vulnerabilities in their child nodes, the root node can be regarded as the integrity test of network host information security. Finally, the network attack chain efficiency model is constructed on the basis of the optimal network host information security integrity test efficiency. Experimental results verify the feasibility and effectiveness of the proposed method.

The adversarial example is an input carefully designed by the attacker to tamper with the output of the neural network model. The emergence of audio adversarial example is to tamper with the end-to-end automatic speech recognition (ASR) system results. Due to the high difficulty of tampering with the ASR recognition results, the noise produced by today’s audio adversarial example generation methods is still easy to be noticed by humans and machines. In this paper, we propose an audio adversarial example generation method that makes audio adversarial example difficult to be perceived based on time-domain restriction. This method hides the adversarial perturbation noise in the speech part of audio by limiting the adversarial perturbation noise in the time domain. Our proposed method is more difficult to be detected by detection methods of audio adversarial example with equal attack performance compared to existing methods. We call the proposed method the HIS (hide in speech) attack.

The biggest feature of P2P system is the direct sharing of resources between users. Its core technology is the positioning mechanism of distributed objects, which is also the key to improve network scalability and solve the swallowing of network bandwidth. In this paper, it uses the HREIK network as the basic P2P platform and lays out the P2P network in a wireless mobile platform. In general, when using mobile P2P networks, people may find some problems, such as poor performance and correctness in pernode lookup. To mitigate these disadvantages, this paper proposes to design a measure for enhancing HREIK (EHREIK) network lookup that can keep the finger table fresh and self-update. In EHREIK, we improved the success rate of HREIK network lookups and the consistency of coverage. At the same time, we reduce the time it takes to find something on a wireless mobile network.

Nowadays, many researchers claim that the self-attention mechanism generates a better performance regardless of circumstances. Having doubted such opinion, we examine different models with scalar dot-product self-attention mechanisms under various experimental settings to develop a comprehensive understanding of such technique. In conclusion, we have verified that the performance of the attention mechanism grows as the dataset input increases in size. In addition, we have observed that our attention layer can impact the model’s performance negatively on small datasets. Moreover, we demonstrate that the attention mechanism is model-dependent: opposite effects may be obtained on the same dataset with different model.

Smart agriculture is a new agricultural production model integrating Internet, cloud computing and Internet of things, which provides precision planting and online services for agricultural production. The accuracy and reliability of agricultural data collection affect agricultural decision-making, so it is very necessary to protect agricultural privacy data. Existing solutions usually require the cooperation of trusted institutions, which is not practical. Blockchain adopts the PBFT consensus algorithm and does not require a trusted third party. Edge computing can greatly improve the efficiency of data aggregation and save data transmission time. Hence, this paper proposes a data collection scheme with a faulttolerant property based on blockchain and edge computing to protect sensitive information in smart agriculture. It is proved to be data security, privacy protection, and fault tolerance with the performance simulation and security analysis.

Token-based replay have become a convention commonly known as a standard method for conformance checking because they are able to calculate the fitness of the model to the behavior contained in the event log. However, it does not automatically handle duplicate tasks. In the case of simultaneous enabled duplicate tasks, the token-based replay method will randomly execute duplicate tasks. If the chosen execution path is not the correct one, the replay of subsequent trajectories may be affected and the process model cannot be effectively evaluated. This research aims to solve the above problems and proposes a fitted evaluation method for duplicate tasks, which executes all possible paths in turn, and at the same time, fitly evaluates the quality of mining models that contain duplicate tasks. The experimental results show that the proposed method can provide a good reference standard for evaluating models with duplicate tasks.

Android malware authors have widely used obfuscation techniques to avoid their harmful apps from detection. In this paper, we present an automatic method for de-obfuscation of Android apps. It is based on Markov models. Our tool mainly focuses on layout obfuscation created by obfuscator ProGuard, which is one of the most widely used obfuscators. The evaluation of the system verifies the effectiveness of the de-obfuscation task.

With the development of social science and technology, the communication process is often accompanied by a variety of encrypted malicious traffic intrusion into the network system. Therefore, it is necessary to build a malicious traffic detection system to solve this problem. This paper proposes a construction scheme of malicious traffic detection system based on machine learning. The malicious traffic data selects the public cic-ids-2017 dataset, and carries out feature engineering processing on the traffic data, which is transformed into a data feature set that is easy to be trained by machine learning model. The performance of various machine learning models are tested and compared. Finally, the optimal random forest model is selected as the malicious traffic detection model. Experiment shows that the model can effectively detect encrypted malicious traffic.

In this paper, we propose an effective and efficient method for constructing 3D mesh for medial axis transform (MAT). Given a MAT approximated as medial mesh, our method generates high-fidelity and watertight 3D triangle mesh even under the exposure of large ill-posed and degenerated medial primitives (medial cones and medial slabs). Experimental results show that our method can produce high-quality 3D meshes and accurately remove degenerate primitives.

Today, with the growth of the electric vehicle market, the demand for power capacity is tight, which has a huge impact on the power grid, aggravates the peak-to-valley difference in power consumption, and seriously affects the normal life of residents. V2G technology is a special smart grid, which can store the excess energy generated in the off-peak power consumption in the battery of the electric vehicle. When the electricity is needed at the peak, the electric energy of the vehicle battery is fed back to the power grid system to maintain the power grid stability, get some benefit from it. This communication process may have loopholes in the security protection of private data. This paper proposes an aggregation mechanism (BA-SAPA) that integrates fog computing and blockchain V2G security authentication privacy data, and designs a power injection data aggregation framework to achieve V2G network security, with the injection of electricity. Introduce relevant algorithms and rules to ensure the security of V2G users' private information.

In order to further improve the execution efficiency of the SM3 cryptographic hash algorithm, give full play to the advantages of the mainstream heterogeneous platform of CPU+GPU, SM3 algorithm message is filled, expanded and iterative compression part on the device end, by reasonably dividing thread blocks to make the full use of thread block resources, use loop unrolling to increase independent memory operation, use shared memory to hide the delay of visiting instructions. Experiments on files of different sizes on NVIDIA Tesla P100, the data show that the SM3 algorithm for GPU achieves an acceleration ratio of 0.593611 to 1.207481 compared with the CPU serial implementation, and the acceleration effect becomes faster with the file size, providing a reference for the implementation of the national secret algorithm on the GPU platform.

To solve the problems of low detection accuracy, long modeling time and slow convergence in intrusion detection system (IDS), an intrusion detection system feature selection method based on improved ant colony algorithm was proposed. In this method, an improved ant colony algorithm is used to optimize irrelevant features in data, and the optimal subset of features is selected by considering three indicators: true positive rate (TPR), false positive rate (FPR) and number of features. Experimental results show that compared with the existing feature selection algorithms, the proposed algorithm has more advantages in reducing the number of features required for robust IDS construction on the premise of ensuring high detection rate and low false positive rate.

As an effective method for the fusion of the digital economy and the real economy, the 3d geometric modeling kernel has received extensive attentions in various fields recently. However, under the rapid development, the safety and quality problems of modeling kernel are constantly occurring, which brings bad user experience to application software developer. Therefore, based on analyzing the specific requirements of modeling kernel function at different modules, we propose a third party test and evaluation method for 3d geometric modeling kernel, which could test and quantify the code security, interface completeness, functional and performance quality level of 3d geometric modeling kernel, so as to assist correct and effective decision-making. Finally, the rationality of 3d geometric modeling kernel test and evaluation method is verified by testing a open source kernel.

Iris recognition is considered as one of the most promising biometrics due to its discriminative features and friendly acquisition methods. Herein, a deep learning-based method is proposed to achieve more accurate and efficient iris recognition. The proposed framework Iris Attention Network (IrisAttenNet) integrates the attention mechanism into a lightweight CNN to extract iris features more specifically. In the process of feature learning, the channel features with more information that contribute to the recognition result will attract more attention and be given higher weights, which is similar to the human visual perception mechanism. The performance of the proposed framework is evaluated by four publicly available datasets representing different intra-class variations: CASIA_Iris_V4 Interval, Lamp, Thousand and UBIRIS.v1. The experimental results have demonstrated that the approach based on the IrisAttenNet shows higher accuracy, stronger generalization and less computational cost. The intermediate outcomes heat maps have proved that the key contribution of the attention module through visualization of the feature areas of images.

CAPTCHA is a key defense technology that improves the security of a website. Text CAPTCHA schemes are the first barrier against attacks on major websites because of simplicity and low cost. In this paper, we proposed a novel, simple, and easy-to-operate feature matching approach to attack text CAPTCHAs. Our method combines a localization network and a feature matching network to attack three real-world text CAPTCHA schemes, including Wikipedia, Baidu and JD. We used 2000 samples for our attack experiments, and the results showed that our proposed method achieved high success rates on all three schemes.

Summarizations to the patients’ problems and the doctors’ treatments in the medical conversations can be highly useful, in terms of helping other patients with similar problems have a precise reference for potential medical solutions. The patient's first question in the medical conversation often contains critical information that determines the diagnosis. The previous simple weight division for the role of dialogue determines that doctors and patients' discourses do not affect each other, which makes it incomplete for the model to obtain the important discourses in the dialogue. Therefore, it is proposed to use episodic memory network to enhance the influence of the patient's first sentence on the model's acquisition of important words, and re-divide the sentence role weight according to the first sentence vector. The experiment proves that using episodic memory network can be more sensitive to the information between utterances, acquire important utterances more accurately, and generate high-quality abstracts.

Using artificial intelligence in automated sleep staging has become very popular. In this paper, we evaluate a deep learning model, Tiny Sleep Net, to illustrate the limitations of the deep learning model in the task of sleep staging predication. During the experiment, we changed the hyperparameter to achieve various study rates, changed the number of epochs and the balance of datasets. The results showed that the actual result of this model is quite acceptable, which is 82.3%. In contrast, the whole model is relatively fragile when some parameters are changed, witnessing an incredibly low accuracy when predicting. In addition, compared with the overall accuracy, the difference in the per-class accuracy can be drastic. Therefore, these results showed that the neural network is susceptible to the hyperparameter and the exact number of epochs generated. Therefore, in exact situations when the input dataset is not balanced or hyperparameters have minus changes, the final prediction accuracy will decrease dramatically.

In order to overcome the demerits of fruit fly optimization algorithm (FOA), such as easily relapsing into local optimum and unstable results which are caused by strong dependence on the selection of algorithm parameters, The cloud model theory is introduced into the algorithm improvement, and the algorithm is optimized and improved from two aspects: the optimal step size of the algorithm and the optimal solution generation mechanism. Firstly, the conception of taste concentration introduced and adjusted adaptively for controlling search step to improve the global search ability and local optimization ability of the algorithm. Then, the randomness and fuzziness of smell concentration parameter is described by normal cloud model and adjusted to finish osphresis search operation automatically to improve the searching precision of the algorithm. Finally, The improved algorithm is used to the automatic test, compared and analyzed with the experiment of other FOA in reference literatures. The results of experiment show that the improved algorithm has better advantages of test efficiency and accuracy.

The traditional hierarchical identity fully homomorphic encryption scheme can only satisfy the homomorphic operation under a single identity, and the dimension of the lattice will increase synchronously with the increase of the layering depth. Aiming at the problem of high algorithm complexity due to the high-precision real number iterative operation used in the trapdoor derivation process of the existing scheme, this paper proposes an improved multi-identity fully homomorphic encryption scheme (mIBFHE): first, we use the features of new MP12 trapdoor to optimize the Gaussian sampling of the original scheme; secondly, the encryption and decryption structure of the original scheme is modified on the basis of the dual Regev system, so that the improved hierarchical identity encryption (HIBE) scheme meets the conversion format of the GSW compiler; finally, the transformed HIBFHE scheme is further transformed into a multi-identity fully homomorphic encryption scheme through the Mask system that meets the conditions. The hardness of our scheme can be reduced to the LWE problem, and IND-ID-CPA security is achieved under the random oracle model. After comparative analysis, the size of the master private key of the mIBFHE scheme in this paper is significantly reduced compared to the size of the same type of IBE scheme, which is 1/4 of the original scheme size. Our scheme is more efficient than traditional IBFHE scheme that it has smaller size than similar schemes in both the public key and the ciphertext when satisfying the attributes of hierarchical, homomorphism, and multiple identities and at the same time.

Cyber security executes a useful function in the information technology sector. Recently, keeping the data in the secured area has grown one of the most rebuts recently. In the cyber security field the first thought that would be concerned is a cybercrime that is regular intensifying extremely. Several states and organizations are getting lots of techniques of avoiding these cybercrimes. Besides several techniques cyber security is still larger impact to the entire community. This research basically persist on rebuts endured by cyber security on current technologies. Also, the orientations modify the image of cyber security and it pursues on current ethics and cyber security techniques.

Static analysis is a famous method to analyze vulnerabilities in software, and value-dependence analysis is an efficient way in static analysis. This paper shows a method to construct value dependency graph for a project by analyzing incremental codes, which can help simplify the process of performing static analysis for a whole project.

The hybrid system is composed of discrete component and continuous component. However, there is no modeling language that can describe the discrete and dynamic characteristics of it at the same time. Hence, we propose a combined modeling method for hybrid system. SysML and Modelica are used to model discrete and physical components of hybrid systems respectively. The structure diagram of SysML and port are used to realize the communication between SysML model and Modelica model, and the complete semi-formal model of the system can be obtained. Since it is impossible to directly make verification on a semi-formal model, we propose a transformation algorithm to transform it into the formal model, and the correctness of the algorithm can be demonstrated by bisimulation. In addition, we select the hybrid interface automata as the formal model for describing the dynamic behavior of a hybrid system and the interface properties between subsystems. In order to obtain the complete formal model of the system, we propose a combination algorithm of HIA based on SysML and Modelica. An automatic air refueling system is used to illustrate the correctness of the combined modeling way and the algorithms proposed in the paper.

As a common illness, atrial fibrillation is difficult to detect, so accurate detection is necessary to ensure better hospital treatment. In this work, we study three topics: the performance of two standard classifiers called Support Vector Machine and Logistic Regression, the impact on the classification result caused by imbalanced data, and which features influence the training model. In evaluating the performance of classifiers, we use four parameters: precision, recall, F1-score, and accuracy. Besides, we also use the ROC curve and the confusion metrics to support our findings. Finally, we get a good classification result and prove that the feature consists of a robust coefficient of variation, and the Skewness parameter shows the most impact on the training model.

In order to ensure that the sink node receives the data sensed by the sensor node according to the reliability requirements expected by the observer, designing a reliable transmission protocol with low power consumption is the primary problem to be solved in data acquisition of WSN. The limited energy of sensor nodes makes it necessary to find a good compromise between reliable transmission and high energy consumption. The protocol proposed in this paper gives good consideration to both reliable data transmission and energy consumption. Simulation experiments and theoretical analysis show that the energy consumption of this protocol is much less than that of HHRA protocol under the same reliability.

There is a problem in Software Defined Network (SDN), is for the controller to select a suitable location, used to achieve the reduction of propagation delay between controller and switch, aiming at this problem, the article puts forward the partition algorithm based on time latency (PABL), the algorithm can guarantee that it can select appropriate location to place controller in each subnetwork, this can reduce the propagation delay between the controller and the switch. Due to the influence of geographical factors, the number of switches in each subnetwork is very different, which would lead to the unbalanced load of controllers among the subnetworks. In order to solve the above problem, this paper proposes a load balancing algorithm based on switch migration (LBABSM). The simulation results show that compared with k-means and k-mediods, PABL algorithm can effectively reduce the propagation delay between controller and switch. After LBABSM algorithm, the load of controller in each subnetwork reaches the equilibrium state.

Voluntary application for college entrance examination is the key link for parents and candidates after college entrance examination. With the development of information technology and the diversification of information access, "Internet+Volunteer Application" has become one of the main ways for students and parents to fill out their volunteer applications. But how to use the huge amount of information on the internet to find suitable colleges and majors for students is an important issue. Therefore, this paper designs an intelligent volunteer application system based on big data and collaborative recommendation technologies to provide intelligent decision-making solutions for candidates to better select universities and majors.

Accurate prediction of residential water consumption is the basis of optimal allocation of urban water resources and an important way to achieve high quality development of the country. In order to improve the accuracy of residential water prediction, aiming at the shortcomings of KNN (K-nearest Neighbor Algorithm) algorithm in water consumption prediction, a water consumption prediction method based on K-means and improved KNN algorithm is proposed. The K-means clustering method was used to determine the categories of historical samples of residential water consumption, and the way of searching similar historical sample sets in KNN algorithm was improved and optimized. The prediction model was built, and the prediction results were compared with those of SVM model and BP neural network model. The results show that the prediction accuracy of the prediction model based on K-means and the improved KNN algorithm is greatly improved, which is feasible and practical, and can be applied to the prediction of residential water consumption.

Digital signatures are widely used in electronic communications, for example in electronic commerce and e-mail. But classical digital signature scheme only offer security relying on computational assumptions. In contrast, quantum digital signatures provide security based on fundamental principles of quantum physics. Quantum group signatures provide anonymity for signatory, quantum blind signatures provide confidentiality for signed messages. In this paper we demonstrate a quantum group blind signature scheme which have the properties both of group signatures and blind signatures. Besides, the security of our scheme is guaranteed by the quantum mechanics, so without any computational assumption. the group property and blind property of our scheme profits from the properties of the measurement-based quantum computation (MBQC). Our quantum group blind signature scheme can be widely used in e-voting system, e-cash, e-commerce, etc.

With the continuous growth of autonomous driving technology, autonomous driving testing puts forward higher requirements in terms of comprehensiveness, authenticity, and efficiency, and traditional road testing can no longer guarantee the testing requirements. The actual traffic scene in my country is relatively complex, which is quite different from the standard traffic scene and data in the simulation software. The simulation scene test cannot meet the accuracy requirements of my country's road test. In view of the characteristics of road testing and simulation testing, this paper proposes an automatic driving test method based on actual road scene data playback. The main test process is to use the vehicle equipped with the automatic driving system to collect the actual road scene, and then classify, identify and save the collected scene, finally complete the function and performance test of the combination of virtual and real automatic driving through the real-time playback of the scene data. The test method uses real scene data for simulation test, which improves the efficiency, repeatability and safety of the test, and meets the high-precision requirements of the test to a certain extent.

Software engineering technology is an important product of computer technology progress and development. Software engineering technology has been widely used in many fields, and has played a good role in promoting the development of social economy nowadays, China has fully entered the era of big data, and the application of big data technology in software engineering is more extensive. The development of social economy provides certain economic support for the progress of science and technology, and software engineering technology has also developed rapidly, showing certain advantages in many fields in the society. The effective combination of software engineering technology and big data drives the equipment and hardware to further promote the level of automation and intelligent development of the society. In view of the combination of big data and software engineering technology, this paper analyzes the key technology of software engineering in the era of big data and analyzes its application to a certain extent.

With the increasing complexity of nuclear reactor design, in order to improve the efficiency and economy of the design and development of complex systems and form a complete digital prototype development system, model-based systems engineering (MBSE) is applied to system design and development. So as to further meet the convenient management of MBSE complex relationship requirements, this paper first analyzes the SysML (systems modeling language) requirements relationship, and builds a requirements management interaction platform based on this, so that it fully conforms to the SysML requirements relationship. Secondly, based on the underlying principle of SysML language, the two-way data binding between the requirements management interactive platform and the modeling software CSM (cameo systems modeler) is realized. Finally, the feasibility of the method is verified by the requirements of the nuclear reactor, and the management of the unified model data and automatic conversion of text to model is realized.

Cell counting is a fundamental study in cell culture research and an essential step in many cell experiments to determine cell numbers and survival rates. This paper uses two different methods to count cells, including machine learning and the deep learning method. It compares the differences, advantages, and disadvantages of the two methods in cell counting through experimental data comparison. In the first method, we use image preprocessing and open-source software to count cell images. The second method is based on CNN cell counting. We input 200, 400, 800, and 1000 pictures for training, selected 268 pictures as the test set, predicted the results, and got a training model with an accuracy of about 91%. Finally, we compared the experimental results and analyzed the pros and cons of the two methods.

Aiming at the problem of optimizing the distribution route of fresh agricultural products, the goal of the shortest total distance, the shortest time, and the highest full load rate is the goal of constructing a fresh agricultural product distribution route optimization model; based on the traditional Saving algorithm, the delivery form and the fast non-dominated sorting strategy are introduced, and an improved saving algorithm for solving the multi-objective vehicle routing problem is proposed, improved Saving algorithm is used to solve the model and conduct empirical analysis. The results show: compared with the traditional saving algorithm, the improved saving algorithm can reduce the total transportation mileage by 23.258%, reduce the delivery time by 21.325%, and improve the vehicle full load rate by 19.339%, the validity of the algorithm and the scientificity of the model are proved.

This paper studies the detection of node data anomalies in the Internet of Things to improve the accuracy of detection. When the content of the external data collected by the sensor nodes in the Internet of Things is random and there are great differences in the characteristics of data storage form and format, it is difficult to achieve consistency between the data of each node. The traditional detection method can not accurately propose the unified features to be detected according to the difference information, resulting in the problem of low detection accuracy. Data monitoring is the basis of the application of the Internet of Things. The current data monitoring methods have the defects of high data repetition rate and low monitoring accuracy. In order to obtain the ideal data monitoring results of the Internet of Things, a traffic prediction algorithm based on quantitative recursive feature extraction is proposed, The time series analysis model of short-term network traffic in network end-to-end routing buffer is constructed. The false nearest neighbor algorithm and average mutual information algorithm are used to reconstruct the phase space of network traffic time series, calculate the percentage of time-frequency feature points in the recursive graph plane to the total number of points in the plane, and realize the time-frequency entropy feature extraction of network traffic, The simulation results show that the algorithm can accurately predict and judge the phase trajectory of network traffic, and the prediction process has good anti-interference ability and high prediction accuracy.

Qingyun Choir is a powerful chorus group in Jiangxi Province. With the development of computer technology and virtual reality technology, the choir can rely on computer technology and virtual reality technology to complete the creation of repertoire, practice and performance. The three virtual synthetic space environment brought by the virtual reality technology allows the audience to see the choir singing in other areas through the projection equipment and the screen. Choir members can contact through the audio recognition in the chorus contact system to ensure that the pitch and pitch of their own chorus can be consistent with other members. Audio workstations can provide music creators with virtual instruments, and music creators can design tracks on the computer to improve efficiency and reduce costs. This paper mainly studies the application of virtual reality technology in Qingyun Choir through three aspects.

As the COVID-19 pandemic spreads across the globe, it highlights the importance of using all available resources to mitigate this common human challenge. Therefore, this paper studies and evaluates the general convolutional neural network and a method based on lung X-ray image classification. Attention mechanism mechanism + DenesNet method for detecting infected patients from chest X-ray images. In order to improve the validity, the mixed data set is preprocessed in this paper. In order to reduce the problem of small number of samples, we adopt transfer learning to transfer the information extracted from the pre-trained model to the model to be trained. The experimental results show that the overall accuracy of the design experiment in this paper has been improved to a certain extent compared with the original network model, and the overall accuracy has reached 83.82%.

The continuously updated and widely applied information technology and internet technology make the amount of data increase with an unimaginable speed. Subsequently, industries have also entered into the “Digital Times”, which brings challenges and opportunities for industrialization. Through industrial big data government technology, we can continuously improve the level of industrial intelligence, the management efficiency and the production efficiency of enterprises, so as to improve the competitiveness of enterprises and the country. However, in the face of this large number of industrial big data with complex and diverse data characteristics, the application of traditional data processing methods in industrial big data government is not applicable, especially in the condition of valuable and scarce computing resources. A new proper data process framework, in this paper, is proposed for industrial data, whose principle is to reasonably select data features by utilizing a max-relevance min-redundancy (mrmr) algorithm that considering the relationship between data characteristics. The effect is verified on a series of industrial datasets. And the experiments results can effectively improve the utilization and analysis efficiency of data, reduce the difficulty of data processing, improve the data storage efficiency and reduce the data storage space.

Machine learning and pattern recognition have been widely used in resting-state functional magnetic resonance imaging (rs-fMRI) data to investigate Alzheimer’s disease (AD). However, many previous methods have not focused on the pre-Alzheimer's disease, mild cognitive impairment (MCI), have mostly classified features from functional separation or functional integration alone, which may have overlooked the correlation between the two. We propose a novel method for MCI diagnosis using the fusion local features of brain area signals as features for functional separation and brain network properties as features for functional integration. The rs-fMRI data of 43 MCI patients and 46 normal cognitive (NC) controls were analyzed using graph theory and nonlinear time series analysis to extract the brain network properties and local cortex signal features. The classifier-SVM achieved an accuracy of 92.8% based on the above features, which is generally higher than those of conventional methods based on functional integration using brain network properties alone or functional separation using local features alone. Our method demonstrates the measure of multi-feature integration of functional integration and functional separation as a powerful tool to classify MCI patients.

Digital image self embedding watermarking technology is one of the main technical means to solve the problem of image tamper detection and recovery. The existing self embedding watermarking technology has the problem of low authentication granularity. In this paper, an image self embedding semi fragile watermarking algorithm based on block truncation coding is proposed. Watermark generation and embedding: the image is divided into two levels at the block level, and the granularity is 4×4 and 2×2 image block; block truncation coding is used for each 4×4 image blocks are compressed and encoded to generate 2×2. Restore the watermark of the image block, and hash the restored watermark for the authentication watermark of the image block. Tamper authentication and recovery: a 4×4. Compare the watermark extracted from the image block with the watermark generated by the block to judge the two of the image block×whether the sub block two passes the authentication; restore the image block according to the authentication result.

Voting-based consensus algorithms such as the Practical Byzantine Fault Tolerance (PBFT) algorithm and Raft algorithm are mainly used in consortium chains. PBFT algorithm is suitable for the Byzantine environment while Raft algorithm is suitable for the non-Byzantine environment. Although there are access mechanisms in the consortium chain, Byzantine nodes may pretend to be honest nodes. PBFT algorithm has high time complexity and poor performance when too many nodes participate in the consensus. Meanwhile, the Raft algorithm can be applied in practice due to its excellent performance and ease of understanding. To solve the performance as well as security problems of the consensus algorithm in consortium chain, this paper proposes the reputation-based Raft algorithm (rbRaft), which can get the honesty degree of the nodes by recording the reputation value of the nodes participating in the consensus. If a node is malicious, its reputation value will be reduced and the node has a low probability to become a leader. If the reputation value is less than 0.5, then it will be removed from the network and will not participate in the consensus. We also add digital signatures to the Raft consensus algorithm, which can guarantee the reliability of the data transmitted between nodes. This paper conducts simulation experiments to evaluate the performance and fault tolerance of the algorithm.

The tokamak plasma experimental device is composed of various subsystems. The data generated by diagnostic systems such as density and electromagnetic measurement are the basis for offline physical analysis to provide data sources for plasma control. The diagnostic data acquisition time varies according to the needs of pulse discharge, and each experiment will generate a large amount of data. With the continuous increase of discharge time, the existing diagnostic database storage system can no longer meet the needs of plasma long-pulse discharge. In order to store long-pulse diagnostic data, it is necessary to design a high-performance diagnostic database storage system with flexible storage mechanism and convenient data retrieval for users. After investigation and comparison, because HDF5's own flexible technology stack can meet the evolving needs of data users, the database uses HDF5 as the data storage format. The overall storage system is based on C/S architecture, and uses TCP/IP technology to collect, transmit, store and display data. The metadata manager is responsible for the mapping of external requests to HDF5 files, and the HDF5 readers and writers are responsible for performing the actual data reading and writing. The storage of the database adopts the time slice method, which realizes fragmented reading, snapshot drawing, and parallel writing. The console interface also records the number of data accesses, and dumps frequently accessed data to a high-speed hard disk. After testing, the system can run stably and has good portability and expansibility.

Based on JSP technology and SSH framework, the online training platform of logistics management provides a large number of training opportunities for current college students majoring in logistics management, changes the current teaching mode of logistics management, and aims to create a simulation curriculum system of teaching practice for students. The online training platform of logistics management relies on network information technology, which can simulate different job roles and help students to understand and get familiar with the specific operations of various business processes in actual logistics management. With functional modules such as dispatching center, warehouse center, transportation center and production management center, a unique "post training mode" is formed, and brand-new teaching methods are organically integrated into the actual logistics management. This strengthens students' hands-on ability and independent innovation ability, improves their familiarity with professional skills, further broadens students' horizons and expands their knowledge fields, so that students can improve their comprehensive quality to adapt to social work through practical training and become a qualified application-oriented high-quality talent.

Based on the Internet and new technology platforms related to smart tourism, including cloud computing platform, cloud service, big data, 3D GIS, mobile Internet and mobile intelligent device technology, Internet of things and other related concepts and applications and based on these advanced technologies and e-government cloud computing service platform system architecture. This paper puts forward the design and research of building the information construction framework of intelligent management cloud platform in Longhu Mountain scenic spot. The specific research contents of this paper include the establishment of a smart Longhu Mountain scenic spot big data center, a smart management cloud platform, and four core functions (the four core business functions mainly include scenic spot comprehensive service function, internal business management function, scenic spot monitoring and analysis function, scenic spot data sharing and exchange function). In addition, this paper studies the system architecture of Longhu Mountain smart scenic spot based on cloud computing service platform to promote the modernization and intelligence of tourism, which not only facilitates regional management and user use, but also realizes a high degree of information sharing and intelligent perception. In short, this research has important theoretical value and practical significance for promoting the integration of tourism and information industry in Longhu Mountain, promoting the development of smart tourism, improving the quality of life of the people, improving the comprehensive development level and the economic development of the people.

In different flow fields, the description and tracking of droplets in computer has high time complexity. In order to improve the efficiency of droplet simulation and calculation, on the domestic DCU acceleration device, the immersed boundary methods is used to realize the simulation of droplet deformation under shear flow. After finite difference of the hot spot pressure Poisson equation, the Jacobian implicit iterative method is used for parallel improvement, divide a larger computing task into multiple independent computing modules, and make reasonable use of the local data share to reduce the number of global memory accesses. The results show that compared to the CPU serial implementation, The heterogeneous droplet deformation simulation achieved the highest speedup ratio of 2.34 under the premise of ensuring accuracy.

In order to ensure the evaluation of carrying capacity of resources and environment and suitability for territorial and spatial development abbreviated as Dual-evaluation work in a scientific, standardized and effective manner, the evaluation system of carrying capacity of resources and environment and suitability for territorial and spatial development is designed and implemented. The system includes Dual-evaluation data resources management, online evaluation, evaluation results output management, guidelines and cases management, and uses GDAL, GeoServer REST service, online evaluation technologies to realize the system data resources storage, online evaluation and evaluation results publish functions. The results show that it has a great significance for to comprehensively promote the Dual-evaluation work for various level users.

The innovation of modern cultural and creative products requires designers to spend a lot of mental work to combine perceptual thinking with rational thinking. Therefore, high-quality product design takes a long time. In order to ensure the diversity and quality of products and effectively improve the design efficiency, a design optimization method for cultural and creative products based on simulated annealing algorithm is proposed. This paper takes the cultural heritage and creative products of Fujian and South Africa as the research object, firstly introduces the value and characteristics of the material cultural heritage of Fujian and South Africa, and then designs the creative products in the material cultural heritage of South Africa based on the simulated annealing algorithm. Geometric topology attribute, using the grid division function to carry out triangular mesh design for intangible cultural heritage products, so as to better guarantee the protection of intangible cultural heritage and design more perfect intangible cultural heritage products.

The application of database programming technology in computer software engineering is mainly for the establishment of the database. The first thing is to improve the database system, start from the network operation environment, build a database establishment framework, reduce the database security risks, and strengthen the confidentiality of data files. In this process, the database programming can clarify the performance of the computer software accordingly, carry out the corresponding computer programming work, and can fill the content in the database structure framework to ensure the resistance performance of the computer software. Therefore, programming technology is the key to the coordination function, enabling the implementation of different modules of computer software.

In the semantic segmentation of images, the problems of incoherence of segmented objects, inconsistency of small objects, and mismatch between objects and labels have become the bottleneck restricting the application of semantic segmentation. Accordingly, an encoder-decoder network is designed with an improved residual network, and an improved model combining attention mechanism and pyramid module is proposed. Among them, the attention module enhances the position information of objects, and the pyramid module improves the segmentation accuracy of object outlines and small objects. The simulation experiments were carried out on the PASCAL VOC 2012 dataset, and the results showed that the MIoU (mean intersection over union, average intersection ratio) reached 79.5% in the improved network structure on the VOC 2012 dataset, which was higher than the accuracy of the basic algorithm. 2.8%, which verifies the effectiveness of the proposed method.