最新刊期
卷 22 ， 期 12 ， 2023
- 摘要：With the continuous development of software technology， the teaching of software engineering courses also needs to fully reflect the latest progress in engineering practice and highlight the cultivation of relevant practical abilities. To this end， explore software engineering teaching methods centered on practical ability， and carry out practical teaching centered on the construction activities of delivering executable software； Simultaneously integrating the cultivation of key competencies in software development and delivery through an incremental and iterative process. In terms of instructional design， course content is arranged according to coding， design， requirements， and other levels， and development projects based on team collaboration are organized. Teaching practice has shown that this teaching method can enable students to experience a relatively complete and complex software system development process， and enable them to understand and master software engineering related knowledge and abilities in a step-by-step and concrete way.关键词：software engineering teaching;practical skills;teamwork project practice
- 98
- |
- 15
- |
- 0
发布时间：2023-12-28 - 摘要：The demand for specialized software professionals in the computer industry is critical in China， prompting many universities， such as Beihang University， to offer software engineering courses to cultivate relevant talent. Given that the demand for software engineering professionals varies across different fields and industries， it is essential to enhance the diversified level and comprehensive quality of software engineering talents. To achieve this goal， many universities have designed software engineering courses with different directions， which not only respond to diversified demands but also mobilize students' personal interests and leverage the professional background of teachers. However， inconsistencies in practice contents and assessment standards remain significant challenges. This article provides an overview of the software engineering course at Beihang University during the spring semester of 2022， outlines the three directions of the software engineering course （Internet software， intelligent software， and embedded software）， and describes their respective practical content and assessment standards. The article analyzes the differences in the score distribution and score indicators of the three directions and conducts theoretical calculations and analyses on the cognitive load and workload of students in the three directions using the NASA Task Load Index. Additionally， the article presents a comprehensive analysis and summary of the student suggestions， highlighting prominent problems and key recommendations. Lastly， the article provides an analysis of the reasons for differences in course practice and assessment standards and proposes feasible solutions to provide guidance and suggestions for the scientific reform of software engineering courses.关键词：software engineering;course practice;assessment methods;consistence;NASA-TLX
- 44
- |
- 6
- |
- 0
发布时间：2023-12-28 -
Research on Software Engineering Teaching Integrating Open Source Software Philosophy and Examples AI导读
摘要：Software engineering is an important discipline in computer science education at the higher education level， aimed at helping students establish a comprehensive understanding of the field of software engineering and providing them with necessary theoretical and practical foundations. Traditional software engineering education mainly involves a combination of classroom lectures， laboratory practices， and project assignments. However， in the rapidly evolving era of software research and development， there may be certain limitations. Therefore， Changsha University of Science and Technology emphasizes the practice of the course "introduction to software engineering" to enhance students' understanding of the open-source software concept. Throughout the theoretical learning of stages such as requirements analysis， software design fundamentals， software architecture design， user interface design， and software detailed design， open-source software cases are integrated. Simultaneously， virtual simulation experiments are introduced to cultivate students' real capabilities in software requirements analysis， allowing them to have a comprehensive and in-depth understanding of software engineering from project requirements analysis， design， development to testing. The focus is on nurturing students' overall software engineering capabilities， leading to comparatively favorable teaching results.关键词：software engineering;open-source software;case study;software requirement analysis;virtual simulation experiment- 33
- |
- 7
- |
- 0
发布时间：2023-12-28 - 摘要：The Excellent Engineer Education and Training Program is a key focus and breakthrough in China's engineering education reform， and practical training is a key focus of teaching in university software engineering programs. Taking the software engineering major at Chongqing University as an example， this paper studies the practical training teaching reform plan under the background of the Excellent Engineer Education and Training Program， and expounds the teaching reform practice in the undergraduate practical training teaching of the software engineering major in the School of Big Data & Software Engineering at Chongqing University， including the practical training practice framework， practical training structure design， practical training process and implementation， teaching services， assessment and evaluation， etc. Through the research， it was found that the new teaching model has improved students' professional ability and teamwork ability， so as to achieve the ultimate goal of cultivating excellent engineers.关键词：software engineering;excellent engineer program;teaching reform;practical training
- 23
- |
- 2
- |
- 0
发布时间：2023-12-28 - 摘要：In order to enhance students' ability to solve complex engineering problems. Aimed at the course objectives of software engineering， a systematic course construction is carried out from the aspects of blended teaching， flipped classroom， course ideological and political education， student learning ability cultivation and course evaluation guided by practice and following the engineering education certification standards. Relevant practices are carried out， which have improved teaching effectiveness and quality effectively.关键词：software engineering;practice-oriented;ability promotion;blended learning;course evaluation
- 12
- |
- 4
- |
- 0
发布时间：2023-12-28 - 摘要：The national strategies on information technology application innovation （ITAI） and emerging engineering education （EEE） have put forward higher requirements to develop information technology engineering talent. To this end， Southwest University has designed a new framework of curricula for innovative software engineering talent education， where a cluster of courses under the background of "ITAI-integrated EEE" are provided. As a core course in this cluster， the reform and exploration of system analysis and design play a crucial role in training the new generation of outstanding software engineers. Considering the challenges such as incorporating ideological education （IE） into curriculum， combining theory with practice， cultivating higher-order abilities， and monitoring the learning process， a new collection of course objectives that adapts to the "ITAI-integrated EEE" is produced， while a "4-NEWs" （i.e.， new-perspective， new-mode， new-approach， and new-technique） classroom is created： Based on the perspectives of "software-defined everything" and "ITAI"， new case studies are extracted from daily events or current events to drive knowledge dissemination and naturally integrate ideological and political education. A new teaching process of "MOOC-Classroom-SPOC- CodeArts" creates online and offline blended （OOB） teaching as well as an integration of course teaching and software production， thereby merging theoretical learning with practical applications.A new pedagogical mode that integrates fun classroom activities with a multidimensional evaluation mechanism is designed to stimulate students' motivation for higher-order learning.A new technique， the facial expressions-based affective computing-enabled intelligent teaching assistant system （iTAS）， is developed and applied for tracking and intervening weak points in OOB teaching activities on time. The practical results indicate that the OBE course objectives of this course is quite good，and it has a noticeable impact on enhancing students' comprehensive problem-solving and software construction abilities.关键词：software engineering education;system analysis and design;information technology application innovation;emerging engineering education;intelligent teaching assistant system
- 23
- |
- 2
- |
- 0
发布时间：2023-12-28 - 摘要：With the development of digital economy， the intelligent audit is replacing the traditional audit gradually. Therefore， this puts forward higher requirements for the training of audit talents. This paper explores an intelligent audit talent training path to train such talents in the development of the digital economy from multiple perspectives， including staff construction， curriculum system， practice platform， effective teaching organization and evaluation methods. The intelligent audit micro-credential serves as a new talent training mode in auditing education， helping to cultivate talents who can quickly adapt to changes in audit work， and provide strong support for the development of the digital economy.关键词：AI audit;micro-credential construction;talent training;curriculum system
- 14
- |
- 2
- |
- 0
发布时间：2023-12-28 - 摘要：In response to the urgent demand for high-quality development of graduate education in the context of Chinese-style modernization drive， this paper aims to explore and deepen the collaborative training model of industry-education integration for graduate students. It addresses the challenge of a shortage of technical transfer professionals in the construction of the national innovation system. The paper analyzes the current situation and issues in the training of technical transfer talents in China. Grounded in the principles of "multiple subjects， interdisciplinary， project-oriented， and compound-type，" the paper conducts practical exploration in the field of technical transfer through graduate education. It establishes a collaborative interdisciplinary training model for graduate students， bridging engineering and business disciplines. This model features a combination of professional degrees and the ability in technical transfer， contributing to the transformation and innovation of the collaborative training model and institutional mechanisms in depth integration of industry and education. The paper aims to enhance graduate students' engineering awareness， practical experience， and innovation capabilities， accumulating valuable practical experience in the process.关键词：interdisciplinary;graduate students;collaborative training;technology transfer;training mode
- 9
- |
- 1
- |
- 0
发布时间：2023-12-28 - 摘要：The mechanism of the integration of industry and education can effectively improve the course quality. Taking the course of "digital image processing and image communication" as an example， this paper discusses how to implement the strategies of the integration of industry and education for specialized courses of electronic information. It proposes the teaching concept of "laying the solid foundation of theory， strengthening the cooperation between universities and enterprises， and improving the quality of students"， and carries out the reform from five aspects： establishing a long-term mechanism for the integration of industry and education， implementing a project-based practical teaching model， building a multilevel textbook system for the theory and applications， conducting multiple evaluation based on process management， and performing continuous improvement to better accomplish the course objectives. It can promote the interaction of knowledge chain， talent chain， and industry chain in the area of image processing. Finally， practical results are demonstrated in terms of the completion status of course objectives， students satisfaction with the course， and the achievements of teachers and students.关键词：integration of industry and education;three-chain linkage;project-based teaching
- 11
- |
- 2
- |
- 0
发布时间：2023-12-28 - 摘要：With the increasingly widespread application of drones， target detection technology in drone application scenarios has important application value and urgent application needs in many fields. In the scene of real-time detection of drone edge devices， due to the presence of a large number of weak target instances in the high-resolution images of drones， the image resolution will be directly reduced， resulting in the loss of weak targets. Therefore， it is crucial to maintain high-resolution information when inputting into the network. To achieve this， high-resolution images are cut into multiple image blocks and used as network inputs to maintain network accuracy while enabling them to run on edge intelligent devices. At the same time， in order to accelerate the model inference speed， group convolution and channel shuffling strategies are used to lightweight design the backbone network of the detection algorithm， and channel attention mechanism is used to improve network accuracy. The experiment shows that compared to YOLOv5， the proposed method improves accuracy by 2% and 4% on the unmanned aerial vehicle dataset VisDrone and the self built dataset OUC-UAV-DET， respectively， and reduces inference speed by 1 ms on the Nvidia hardware （Xavier）. At the network pruning level， combining specific datasets for channel pruning of detection models can reduce inference speed by 2 ms while maintaining algorithm accuracy. In addition， for the single category detection task of unmanned aerial vehicles， optimizing the output part based on the relatively fixed size of the target instance can reduce the number of model parameters by 30%， resulting in a maximum reduction of 2 ms in inference speed.关键词：unmanned aerial vehicle;high resolution images;object detection;lightweight network;network pruning;model deployment
- 36
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：In the field of education， the knowledge system in the discipline is usually linked to construct the discipline knowledge graph. However， because the knowledge system is not immutable， it is particularly important to update the discipline knowledge graph. As a basic technology of building and updating knowledge graph， entity link is very critical. Therefore， in order to update the multi-modal knowledge graph in computer field， the paper proposed a visual entity linkage method. Firstly， we used YOLOV5 to identify the visual entities in the image. By combining visual entities， the triples in image text description were extracted and filtered. Secondly， we proposed a visual entity link rule （VELR） for the computer field. In order to effectively complete the link， the image visual entity and the head and tail entity of the filtered triplet were used successively to find the best link point. We set different link strategies for different link points. The rules made full use of image visual entities and filtered triples to expand the linked entities. Finally， the triplet classification experiment shows that the proposed method is correct to realize the triplet generated by the linkage of multi-modal knowledge graph in the field of computer science， which proves the validity of the proposed method.关键词：computer science;multi-modal knowledge graph;entity linking;visual entity;text description
- 15
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：At present， most classification algorithms for imbalanced data streams require a large amount of historical data to be saved and repeatedly scanned during the training process to improve classification accuracy， which is inconsistent with the single channel characteristics of data streams， and the infinite nature of data streams requires a lot of memory space consumption. To this end， a classification method for imbalanced data streams based on integrated undersampling and online sequence exceeding learning machine （EU-OS-ELM） is proposed. Firstly， the base classifier selects an online learning algorithm OS-ELM that is suitable for data streams； Then， using non replacement random undersampling to construct a training set to enhance the robustness of the algorithm； Finally， a fixed size matrix is used to store the feature information of historical data， which improves the accuracy of data stream classification while minimizing the required additional memory space. The comparative experimental results between EU-OS-ELM and mainstream algorithms on some datasets show that the proposed algorithm only requires an additional memory space of 0.8906 KB on all datasets， proving the effectiveness of the algorithm.关键词：imbalanced data stream;OS-ELM;classification;storage;ensemble
- 12
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：Weakly supervised action localization detect the temporal boundaries of action instances and identify their corresponding action categories with only video-level labels. Due to the lack of frame-level classification labels， weakly-supervised action localization has the problems that some action frames with inconspicuous features are difficult to identify and action frames as well as context frames in videos are easily confused. To address both problems， a weakly supervised action localization method based on attention mechanism context modeling is proposed. This method added semi-soft attention on the basis of action-backgroud attention for guiding the model to focus on frames with insignificant action features； To separate action frames and context frames ， our method utilized context attention for modeling video contextual information. The experimental results show that our proposed method has better action localization effect. When the IoU （Intersection over Union） value is 0.5， the average detection accuracy （mAP） on the THUMOS14 and ActivityNet1.3 public datasets reach 32.6% and 38.6% respectively， which is better than existing weakly supervised action localization models.关键词：weakly supervised;action localization;attention mechanism;semi-soft attention;context modeling
- 26
- |
- 2
- |
- 0
发布时间：2023-12-28 - 摘要：Aiming at the problems of weak generalization， low recognition rate， large amount of parameters and low quality of feature extraction of traditional neural network， a new model structure named STN-GResnet is proposed in this paper.Using part of the Resnet18 network structure and adding the Ghostmodule to enhance convolution features.Combined with the spatial transformation network， the extracted features are space invariant.Meanwhile，the additional angle margin loss function （ArcFace） is added to this network model to train and optimize loss.The difference of texture and granularity of ultrasonic liver CT image is enhanced by optimizing the characteristics of categories.The pre training parameters of transfer learning are used， and then the sample set in ultrasonic liver CT images is improved through data enhancement to avoid the phenomenon of model over fitting caused by insufficient sample size. Finally， the objective recognition rate of the model is 95.7%.What's more，the model is small and the operation efficiency is high.关键词：deep learning;medical image recognition;transfer learning;data enhancement
- 26
- |
- 1
- |
- 0
发布时间：2023-12-28 - 摘要：The traditional sports testing method needs professional testing equipment and personnel， which has the problems of high labor cost and inaccurate manual counting.Therefore， the purpose of this paper is to study a set of intelligent sports test system to help colleges and universities to get more accurate and more efficient sports test results.In order to improve the recognition accuracy， we proposed a model of posture recognition of spatial fusion（Pos-RSF）based on the improved OpenPose algorithm， which could realize the fast and accurate recognition of human motion posture estimation in sports tests and improved the operating efficiency of the model. The experiment showed that the posture recognition model we studied has higher accuracy and better real-time， the overall score PCKh@0.2 was 91.2%， a significant improvement over the most advanced technology， and can meet the basic requirements of sports testing.关键词：sports testing;human motion posture;real-time recognition;posture features;image recognition
- 12
- |
- 2
- |
- 0
发布时间：2023-12-28 - 摘要：In order to solve the problems of traditional collaborative filtering algorithms （CF） that cannot highlight personalization and have low user reliability and time correlation， a hybrid collaborative filtering algorithm based on dynamic time weighting （HCFADT） is proposed. Firstly， for users and projects， calculate the number of target user comments and user comments during the observation time， and introduce dynamic time weights into the project based and user based algorithms. Then， combine the two algorithms and adjust the degree of dependence of the two algorithms by adjusting the factors. The experimental results show that compared to traditional collaborative filtering algorithms and algorithms based on dynamic time weights， the hybrid algorithm has smaller fluctuations， and the deviation of average recommendation prediction is reduced by 3% and 2%， respectively.关键词：collaborative filtering;time weight;hybrid algorithm;recommendation algorithm
- 18
- |
- 1
- |
- 0
发布时间：2023-12-28 - 摘要：At present， online teaching is becoming more and more common in the teaching of information and communication technology courses in higher vocational colleges， and the prediction of students' online learning performance is becoming more and more important. Although teachers lack the advantage of supervision of offline teaching， they benefit from the abundant teaching data provided by online teaching platforms， which can help teachers grasp students' learning situation to a greater extent. The analysis of objective data is of great significance to the early warning of students with poor learning results and to the improvement of students' online learning results. In the online teaching process of ICT courses， the performance indicators of the existing model and the existing model are compared and analyzed， and a novel prediction model of students' online learning achievement is proposed based on voting. The experimental results show that the performance of this model is better than other existing prediction models.关键词：information and communication technology courses;online teaching;performance prediction model;voting model
- 22
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：There are multiple evacuation paths in densely populated areas with multiple decision points. Reasonable planning of evacuation paths can avoid long-term local congestion and reduce the probability of stampede accidents. The local fastest algorithm of the personnel emergency evacuation simulation engineering software Pathfinder cannot solve the problem of uneven path utilization in multi-step decision-making places.Therefore， a path optimization method with the goal of minimizing the total evacuation time is proposed. A feasible path model and path optimization model are constructed based on decision points， and the optimization model is solved based on the Bellman principle to obtain the optimal path plan. According to this plan， the evacuation process is deduced， and the optimal path is updated until no better path appears multiple times， ultimately determining the optimal path and its evacuation time. Taking the evacuation of a student apartment as an example to demonstrate the optimization effect of this scheme， the results show that compared with the Pathfinder simulation results， the total evacuation time of this method is reduced from 330.3 seconds to 292.3 seconds， the utilization time difference of the first floor staircase exits on the east and west sides is reduced from 75 seconds to 1.5 seconds， and the number of people passing through is reduced from 138 to 0 people. This path optimization method can be used to improve evacuation simulation software.关键词：crowded places;multi-decision nodes;evacuation path optimization;Bellman principle;Pathfinder
- 12
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：To improve the measurement accuracy of cerebral oximetry devices using near-infrared spectroscopy， it is necessary to select the optimal source-detector （SD） distance. Conventional cerebral oximetry devices with a fixed SD distance for the sensor probe part are mostly selected based on experience and without specific theoretical studies， which can cause inaccurate measurement results in actual use. According to the transmission characteristics of light in the head tissue， a set of characteristic parameters of cranial tissues is selected， and the five-layer cranial brain tissue model is simulated using Monte Carlo method to study the transmission characteristics of light in the head tissue， and the maximum reachable depth and scattered transmission distance of photon transmission in the tissue are analyzed according to the transmission path of photons. By analyzing the effective depth ratio （EDR） and gray matter layer information ratio （GMLIR）， a set of optimal distances SD=1 cm and SD=3.5 cm are selected to fit this group of selected cranial parameters. Compared with the traditional detection device， by optimizing the distance between the light source-detector， it will provide a new idea for the sensor design of the cerebral blood oxygen detection device and the measurement accuracy improvement of the device.关键词：cerebral blood oxygen detection;light source-detector distance;sensor;Monte Carlo method;tissue model
- 13
- |
- 1
- |
- 0
发布时间：2023-12-28 - 摘要：Nowadays， the traditional mode with single database has difficulty meeting with the demand of in diversified data management. The integration of multi-mode heterogeneous databases has become a research hotspot for overall control and collaborative operation of the global database system. Aiming at this problem， this paper studies the SQL parsing and transformation methods for heterogeneous databases. And interoperability between different database has been achieved through universal intermediate-representation-model establishing， syntax tree parsing， semantic analysis and model transformation. In the functional test based on the TPC-H benchmark dataset， the frame-based test system has 100% support for data types and syntax operations， while the framework has advantages over official tools in terms of operation speed for cross platform addition， deletion， modification， and query， with 13.1 ms， 8.8 ms， 22.5 ms， and 2.3 ms， respectively. The experiment verifies the correctness and feasibility of the proposed method.关键词：heterogeneous database;intermediate representation;syntax parsing;grammar transformation
- 25
- |
- 1
- |
- 0
发布时间：2023-12-28 - 摘要：Particle swarm optimization （PSO） algorithm has been widely used in large-scale complex problems such as resource allocation in recent years because of its simple implementation and easy operation. However， the slow convergence speed and low solution accuracy of the algorithm also restrict its further application. In order to solve the above problems， this paper introduces the chromosome crossing characteristics of genetic algorithm， and combines with comprehensive learning particle swarm optimization， proposes a comprehensive learning particle swarm optimization based on optimal crossing. It can improve the convergence speed of the algorithm and the accuracy of solving the problem by performing the optimal crossover operation between the global optimal particle position and the historical optimal position of the individual to obtain a better individual. The experimental results of benchmark function show that the proposed algorithm has faster convergence speed and optimization accuracy than the original algorithm， and the results of Friedman test and Wilcoxon signed-rank test show that the proposed algorithm has better advantages than other comparison algorithms.关键词：particle swarm optimization;genetic algorithm;comprehensive learning strategy;optimal crossover;Friedman test;Wilcoxon signed-rank test
- 10
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：As a new method of software testing， crowdtesting is mainly to configure scattered public resources to complete specific software testing tasks after receiving tasks online， in order to find software defects that cannot be found or completed by traditional testing. This paper mainly studies the problems in the process of crowd test， such as uneven ability levels of the crowd test workers， inaccurate classification of the ranks of the crowd test workers， and low enthusiasm of the crowd test workers in practical projects. Firstly， the factors that affect the ability level of the workers in the crowd test and the enthusiasm of participating in the task are analyzed respectively， and the standard and method of calculating the ability level of the workers in the crowd test and the model of effective management of the workers in the crowd test are proposed. Then， based on the calculated ability level value of the crowd test workers， the workers' grades are divided， so as to better manage the crowd test workers and test tasks， and improve the enthusiasm of the crowd test workers to participate in the test tasks. Based on the personnel and task data of 50 real test tasks of the crowd intelligence platform， this paper analyzes the data by using SPSS tool， and applies the analysis results to the proposed method. The results show that it can effectively solve the calculation of the ability level value of workers， and improve the enthusiasm of workers to participate in the task.关键词：crowdsourcing test;data-driven;software test;people management
- 12
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：Web-based questionnaires， with their simplicity， speed and cross-territory advantages， provide a quicker way for education and research institutions to obtain information about teachers. In order to obtain accurate and objective evaluation results， it is necessary to design corresponding questions to cover different groups of subjects in different situations， forming a huge reserve of questions. The existing web-based survey model uses a full question bank of assessment regardless of role， coupled with a multi-disciplinary joint survey， which can easily cause subject fatigue and reduce survey efficiency， hindering the large-scale application of web-based questionnaire education monitoring. It is a better solution to group subjects dynamically according to their responses and to distribute targeted question selection. This study proposes an implementation method based on the Angular framework， which supports multi-combination logical jumping and multi-question joint checking. It can automatically make logical judgments and dynamically jump to subsequent groups of questions based on subjects' existing answer combinations during the test， and add the function of joint constraint checking for multiple questions with logical relationships， improving the relevance， effectiveness and accuracy of the assessment and saving test time. The method has been well applied in the Chongqing basic education monitoring teacher questionnaire in 2020 and 2022.关键词：Angular;MongoDB;questionnaire;topic management by dimension;multi-combination logic jumping
- 28
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：In response to the call of the country to master core technologies and accelerate the promotion of domestically produced autonomous and controllable substitution， we provide a teaching environment for domestically produced computing courses to teachers and students on campus. Shanghai Jiao Tong University is the first among domestic universities to build a high-performance computing platform based on the Huawei Kunpeng processor， providing services to a wide range of scientific computing users and teachers and students on campus. The platform is built using the Huawei Kunpeng 920 processor， and is configured with the same SLURM job scheduling system and a shared unified Lustre parallel file system as the π 2.0 supercomputing platform to ensure a consistent user experience. By using two software management methods， Spark compilation deployment and Singularity image deployment， the platform's list of scientific computing software has been enriched. Common scientific computing software such as LAMMPS and GATK have been tested， proving the platform's superior availability and performance indicators. At the same time， a computing course teaching support environment has been constructed based on the platform， which can support mixed online and offline teaching and large-scale virtual simulation experimental teaching. The construction of the platform has accumulated experience for the construction and use of ARM supercomputing in brother universities， improved teaching quality through practical teaching support， and provided better conditions and opportunities for students' learning and development.关键词：high performance computing;Huawei Kunpeng;ARM architecture;performance testing;course teaching
- 12
- |
- 2
- |
- 0
发布时间：2023-12-28 - 摘要：A spectrum acquisition and driving system based on InGaAs image sensor and FPGA control was designed for near-infrared Raman spectroscopy detection. The difference between this system and a single channel output sensor is that the output feature is odd even interleaved dual channel signal output， which uses FPGA to complete the driving control of the internal timing of the system； Simulate the output signal of the system through ADC analog-to-digital conversion chip and perform collection and conversion， and store and sort the data through Dpram module； Real time transmission of spectral data through serial port. The experiment shows that the system has good collection performance， adjustable integration time， high integration of hardware circuit and software， and has high application value in near-infrared Raman spectroscopy collection systems.关键词：InGaAs image sensor;odd-even staggered signals;analog to digital conversion;FPGA;Dpram
- 10
- |
- 2
- |
- 0
发布时间：2023-12-28 -
Study on the Influence of Heuristic and Systematic Clues on the Susceptibility of Spear Phishing AI导读
摘要：Spear phishing emails are highly targeted and deceptive， and their high susceptibility deserves attention. Based on a heuristic systematic model， functional near-infrared spectroscopy technology is used to explore the impact of heuristic and systematic cues on the susceptibility of spear phishing. The analysis of cognitive neural experimental data shows that both heuristic and systematic cues lead to a higher susceptibility to spear phishing. Compared to heuristic cues， individuals invest more cognitive effort in facing systematic cues， and there is no significant correlation between cognitive effort and susceptibility to spear phishing. The research results provide a reference for the study of the susceptibility impact mechanism and prevention practices of spear phishing.关键词：spear phishing;susceptibility;heuristic-systematic model;functional near-infrared spectroscopy- 10
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：Aiming at the problems of scarcity of real samples of labels， low accuracy and poor generalization ability in network intrusion detection， a migration learning network intrusion detection method based on ICNN-MMD combined with inter-domain distance is proposed to achieve domain adaptation and feature migration by reducing the distribution difference between source domain and target domain. First， the improved convolutional neural network （ICNN） is used to pre-train the optimal model on the source domain data； Then， the maximum mean difference （MMD） is used to measure the distance between the source domain and the target domain， and the secondary training is conducted in an unsupervised way to adapt to the domain and share the network weight. The experiment on NSL-KDD dataset shows that the accuracy of the proposed model in this paper is 1.5% and 2.96% higher than that of ICNN in KDDTest+ and KDDTest-21， respectively， which improves the overall index of the network intrusion detection model and provides a new method for detecting network intrusion.关键词：network intrusion detection;transfer learning;ICNN-MMD;domain adaption;network security
- 21
- |
- 2
- |
- 0
发布时间：2023-12-28 - 摘要：Code distribution technology is a key technology for later software maintenance of wireless sensor networks. In response to its susceptibility to interference， a fountain code based code distribution protocol adaptive mechanism is proposed. Theoretical analysis and simulation are conducted to obtain the optimal redundant length that minimizes data traffic under a given packet loss rate； Based on the real-time estimated link quality， dynamic adjustment of the number of redundant packets and replacement of data channels are used to reduce total communication volume and latency. The experimental results show that compared with fixed redundancy length code distribution protocols， DRD with dynamic redundancy length mechanism can reduce communication traffic by about 20.45% and distribution time by 37.39% when the packet loss rate is less than 0.1； When the packet loss rate is between 0.1 and 0.25， the distribution protocol DRD&MCN combining multi-channel and dynamic redundancy length mechanism can reduce approximately 53.08% of communication traffic and 64.05% of distribution time. The proposed method can significantly reduce communication volume and distribution time， improve distribution efficiency， and is beneficial for extending the survival time of wireless sensor networks.关键词：wireless sensor network;code dissemination protocol;dynamic redundancy length;multi-channel
- 11
- |
- 1
- |
- 0
发布时间：2023-12-28 - 摘要：The types and forms of malicious attacks are constantly changing， and the number of attacks is gradually increasing. Traditional neural network model architecture plays an important role in improving model accuracy， reducing model computation and improving reasoning speed， etc. However， traditional model architecture requires a lot of computing resources in search， and its generalization ability is not high. In this regard， it is necessary to propose solutions for network attacks in the context of big data. Based on the application of deep learning in network security， combined with principal component analysis （PCA） and deep Extreme Learning Machine （DELM） in the field of intrusion detection， a lightweight neural network PCA-DELM is designed to reduce computing resources and improve generalization ability while retaining the advantages of traditional neural network model architecture. The simulation results show that compared with other algorithms， the optimized lightweight neural network model PCA-DELM can significantly improve the ability of intrusion detection and speed up the detection rate on different data sets.关键词：intrusion detection;network security;extreme learning machine;principal component analysis;deep learning
- 18
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：Edge perception image processing is an important topic in the field of computer graphics， with many applications in image detail enhancement and tone mapping. However， existing processing methods are often affected by halo and gradient reversal artifacts. To this end， a novel mapping function is proposed， which can flexibly process image details while maintaining edges， making it suitable for various applications. Simultaneously， a reconstruction model based on gradient domain regularization is proposed for edge perception image processing. This model can significantly reduce the impact of halo and gradient reversal artifacts by reconstructing the processed image from the mapped gradient； And this model can be quickly and effectively solved based on Fourier transform， with a processing time of 0.46 seconds for a 1 megapixel image on the Intel i7-6700 CPU. The qualitative and quantitative experimental results show that the image processing method has achieved good results in image enhancement and tone mapping， with detail enhancement indicators SSEQ of 11.39 and ILNIQE of 27.52； The tone mapping indicators SSEQ are 14.88 and ILNIQE are 22.78.关键词：edge-aware image processing;mapping function; regularization
- 11
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：To address the difficulty in identifying detailed features of images captured by visible light imaging devices， a fuzzy cropping contrast limited adaptive histogram equalization （ADFS-CLAHE-FC） image enhancement technique is proposed by combining principal component analysis and an improved anisotropic diffusion filter with two different methods to extract meaningful information from the image. Firstly， the image is denoised using PCA， and then the denoised image is enhanced using ADFS-CLAHE-FC. Finally， the noise of the image is further reduced based on ADFS-CLAHE-FC， maintaining contrast and brightness. Experiments have shown that this method enhances image contrast while eliminating image noise， resulting in better visual performance. Compared to histogram equalization （HE）， contrast limited adaptive histogram equalization （CLAHE）， and other methods， it performs better in improving image quality and preserving image details， which helps to improve the accuracy of image segmentation and extraction.关键词：image enhancement;CLAHE;principal component analysis;contrast enhancement;histogram equalization
- 12
- |
- 1
- |
- 0
发布时间：2023-12-28 - 摘要：Aiming at the problems of difficult partial segmentation detection， many background interferences and easy neglect of small lesions in new coronary pneumonia CT films， a segmentation method based on attention mechanism to improve U-Transformer is proposed. The attention mechanism is used to increase the accuracy of segmentation， and the attention module in the middle of the convolutional layer of the U-Transformer network is modified， and the cross-attention mechanism is used to realize the network to segment the lesion edge more accurately. The whole-local segmentation strategy is added to the network structure to achieve more accurate extraction of small lesion points. The experimental results show that the improved method improves the accuracy by 5.96%， the recall rate by 7.11%， and the sample similarity by 6.49% compared to the U-Transformer， indicating that the improved method has a good effect on extracting small lesion points. Expanding deep learning methods to medical imaging diagnosis can help radiologists diagnose conditions more quickly and effectively.关键词：Covid-19;image segmentation;U-Transformer;attention mechanism;global-local segmentation strategy
- 18
- |
- 3
- |
- 0
发布时间：2023-12-28 -
Efficient Tree Structured Point Cloud Generative Network Based on Probability Distribution Learning AI导读
摘要：Aiming at the non-uniformity of point cloud generation and the low efficiency of model training by TreeGAN （Tree Generative Adversarial Networks）， an efficient tree structure generative adversarial network probability-TreeGAN based on probability distribution learning of point clouds is designed to improve the accuracy of point cloud position. The method learns the probability distribution information of the point cloud and the feature information of the point cloud respectively during the model training， and changes the position distribution of the point cloud by mapping the probability distribution learned in the network.The generated point cloud is well distributed and have good structure.The local features of the generated point cloud can be changed by latent code. According to the experimental results on the ShapeNet Part data set， the method is better than the previous method in the evaluation of JSD （Jensen-Shannon Divergence） indicators， including aircraft， chair single category and overall 16 categories. Compared with traditional 3D point cloud generation models， our method can generate more uniform point cloud， and the model training is more efficient.关键词：point cloud;generative adversarial network;probability distribution;tree structure generating network;mapping network- 13
- |
- 1
- |
- 0
发布时间：2023-12-28 - 摘要：In order to solve the problems of the traditional point cloud simplification algorithm， such as easy feature loss and more holes， a point cloud simplification algorithm based on neighborhood normal angle and dichotomous K-means clustering feature extraction is proposed. The algorithm uses the normal angle of the point cloud neighborhood to extract the whole feature points of the model. Using point cloud curvature as the clustering feature of dichotomous K-means clustering algorithm， the local feature points of the model were initially extracted. Then， the average curvature of the initial local feature points is calculated， and the points larger than the average curvature are extracted as the final local feature points， and the feature of the model is retained. For the non-feature points， the improved voxel reduction method is used for downsampling. The extracted global feature points， local feature points and non-feature points are combined to complete the simplification of point cloud. Compared with the traditional reduction method and other methods in literature， the proposed algorithm has higher precision.关键词：point cloud reduction;feature extraction;normal included angle of neighborhood;dichotomous K-means clustering;point cloud curvature
- 15
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：To address the issue of the susceptibility of insect pest and sticky insect board images collected on site in greenhouses to noise such as uneven lighting and light reflection， which affects recognition accuracy， an intelligent recognition system for greenhouse pests is built based on image segmentation algorithms and linear support vector machines. This system utilizes the fusion of Prewitt edge detection binary image and Canny edge detection binary image with global threshold segmentation to achieve precise segmentation of pest areas in sticky insect board images. Artificial construction of a linear support vector machine training dataset based on segmented pest images， and expansion of single sample feature extraction to 12 based on planthopper and thrip features； Construct a linear support vector machine based pest recognition classifier based on the expanded training dataset， and use gradient descent method for model training. The experimental results show that this method can quickly and accurately segment the pest target area in the sticky insect board image， with an average accuracy of 96.3%； For segmented images， this method can accurately identify pests with an average accuracy of 96.1%， with recognition accuracy rates for planthoppers and thrips being 95.4% and 96.8%， respectively.关键词：edge detection;support vector machines;greenhouse pest;pest identification
- 13
- |
- 2
- |
- 0
发布时间：2023-12-28 - 摘要：In order to ensure the security of data files and user privacy， the most immediate protection is achieved by encrypting data files， which only authorized users can use their decryption keys to obtain the desired content. However， traitors or malicious authorized users may change the original data and leak their decryption keys or directly participate in the construction of the black-box， resulting in illegal data tampering and access. Therefore， it is very necessary to discover and find traitors. Ensuring the authenticity and traceability of the traitor's identity through technical means is one of the important research directions of attribute-based encryption. Based on the overview and security research of traceable attribute-based encryption schemes at home and abroad， the basic concepts and models of traitor tracking are introduced. The tracking process， advantages and disadvantages， tracking costs， security analysis and other issues of the tracking scheme are deeply researched and discussed. Finally， the overall overview of traitor tracking technology based on attribute encryption is summarized， and the existing problems at this stage and possible future research development directions are proposed.关键词：traitor tracing;attribute-based encryption;black-box tracing;white-box tracing
- 15
- |
- 0
- |
- 0
发布时间：2023-12-28 - 摘要：The recommendation of legal provisions is a key task in judgment prediction， aiming to predict the legal provisions involved in the case based on the description of the legal documents and the facts. Traditional machine learning based methods are difficult to extract deep features of case descriptions， and in recent years， deep learning based methods have achieved good recommendation results. Systematically review existing research results from multiple perspectives such as the construction mode， deep learning methods， and complex application scenarios of Chinese legal article recommendation models， with a focus on using convolutional neural networks， recurrent neural networks， graph neural networks， and hybrid models as indexes to analyze their applications in Chinese legal article recommendation. Compare the performance of deep learning algorithms for Chinese legal article recommendation on public datasets. By analyzing existing research， it is concluded that the construction of datasets， complex scenarios， and interpretability issues will be the challenges faced by future research on the recommendation of Chinese legal provisions.关键词：Chinese law articles recommendation;deep learning;judgment prediction;text representation
- 18
- |
- 0
- |
- 0
发布时间：2023-12-28
- Technical support is provided by Beijing Founder electronics co., LTD 鄂ICP备12016542号-70
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0