Multi-feature fusion: Graph neural network and CNN combining for hyperspectral image classification

MDEU-Net: Medical Image Segmentation Network Based on Multi-Head Multi-Scale Cross-Axis

Significant advances have been made in the application of attention mechanisms to medical image segmentation, and these advances are notably driven by the development of the cross-axis attention mechanism. However, challenges remain in handling complex images, particularly in multi-scale feature extraction and fine-detail capture. To address these limitations, this paper presents a novel network architecture, multi-head multi-scale cross-axis attention MDEU-Net, that leverages a multi-head attention mechanism processing input features in parallel. The proposed architecture enables the model to focus on both local and global information while capturing features at various spatial scales. Additionally, a gated attention mechanism facilitates efficient feature fusion by selectively emphasizing key features rather than relying on simple concatenation and improves the model's ability to capture critical details at multiple scales. Furthermore, the incorporation of residual connections further mitigates the gradient vanishing problem by enhancing the model's capacity to capture complex structures and fine details. This approach accelerates computation and enhances processing efficiency, while experimental results demonstrate that the proposed network outperforms traditional architectures in terms of performance.

Optical Fiber Vibration Signal Recognition Based on the EMD Algorithm and CNN-LSTM

Accurately identifying optical fiber vibration signals is crucial for ensuring the proper operation of optical fiber perimeter security warning systems. To enhance the recognition accuracy of intrusion events detected by the distributed acoustic sensing system (DAS) based on phase-sensitive optical time-domain reflectometer (φ-OTDR) technology, we propose an identification method that combines empirical mode decomposition (EMD) with convolutional neural networks (CNNs) and long short-term memory (LSTM) networks. First, the EMD algorithm decomposes the collected original optical fiber vibration signal into several intrinsic mode functions (IMFs), and the correlation coefficient between each IMF and the original signal is calculated. The signal is then reconstructed by selecting effective IMF components based on a suitable threshold. This reconstructed signal serves as the input for the network. CNN is used to extract time-series features from the vibration signal and LSTM is employed to classify the reconstructed signal. Experimental results demonstrate that this method effectively identifies three different types of vibration signals collected from a real-world environment, achieving a recognition accuracy of 97.3% for intrusion signals. This method successfully addresses the challenge of φ-OTDR pattern recognition and provides valuable insights for the development of practical engineering products.

Polarization of road target detection under complex weather conditions

Polarization imaging technology can be applied to unveil the interaction between light and matter by harnessing the transverse vector wave attributes of light, thus to accentuate target characteristics amidst complex weather conditions. This technology has the potential to be widely used in road target detection. However, polarization detection is significantly affected by illumination and detection angles, as well as the considerable variation in the scale of road targets. The optimal polarization parameters should be adaptively adjusted to weather conditions, angles and target features, whereas most existing research employs handcrafted polarization parameters without considering actual complex detection requirements, which are unable to adaptively adjust the polarization feature enhancement methods. In this paper, we propose a road target detection algorithm based on an end-to-end adaptive polarization coding method, named YOLO-Polarization of Road Target Detection (YOLO-PRTD). To enhance the polarized features of targets under complex weather conditions, an Adaptive Polarization Coding Module (APCM) is designed. This module integrates channel-wise global self-attention and small kernel convolution to adaptively adjust the polarization enhancement method using dynamically extracted global and local polarization feature information. A multi-scale detection network is also designed to fully extract and fuse multi-scale feature information from receptive fields, channels, and spaces in different dimensions. Additionally, a dataset of Polarized Images of Road Targets in Complex Weather conditions (PIRT-CW) is proposed for training and evaluation. Experimental results on the PIRT-CW show that the YOLO-PRTD algorithm achieves a mAP0.5 of 89.83%, reducing the error rate by 15.54% compared to the baseline network YOLOX.

A Joint Network of Edge-Aware and Spectral-Spatial Feature Learning for Hyperspectral Image Classification

Hyperspectral image (HSI) classification is a vital part of the HSI application field. Since HSIs contain rich spectral information, it is a major challenge to effectively extract deep representation features. In existing methods, although edge data augmentation is used to strengthen the edge representation, a large amount of high-frequency noise is also introduced at the edges. In addition, the importance of different spectra for classification decisions has not been emphasized. Responding to the above challenges, we propose an edge-aware and spectral-spatial feature learning network (ESSN). ESSN contains an edge feature augment block and a spectral-spatial feature extraction block. Firstly, in the edge feature augment block, the edges of the image are sensed, and the edge features of different spectral bands are adaptively strengthened. Then, in the spectral-spatial feature extraction block, the weights of different spectra are adaptively adjusted, and more comprehensive depth representation features are extracted on this basis. Extensive experiments on three publicly available hyperspectral datasets have been conducted, and the experimental results indicate that the proposed method has higher accuracy and immunity to interference compared to state-of-the-art (SOTA) method.

Optimizing energy storage plant discrete system dynamics analysis with graph convolutional networks

Addressing the challenges of suboptimal model performance and excessive parameters and operations in the optimization of energy storage power plants utilizing Graph Convolutional Network (GCN), this paper introduces a novel approach - the packet-switched graph convolutional network. Initially, a GCN extreme learning machine is established. Drawing inspiration from this solid foundation, we have innovatively crafted a group exchange graph convolution module. This module leverages group graph convolution techniques to amalgamate unique node feature information, tailored to diverse topology graph matrices based on various groupings. This innovative approach ensures that information flows freely and effectively among distinct groupings. Furthermore, we have designed a cutting-edge timing depth separation convolution module, comprising two innovative components. The first component introduces timing depth separation convolution, revolutionizing the original timing convolution module. The second component, the packet-switching graph convolutional network, revolutionizes the time sequence depth separation convolution process. It achieves this by employing 1 × 1 convolutional layers between different feature fusion packets, enabling seamless information exchange between distinct packets. Experimental results demonstrate the efficacy of the proposed model, with root mean square error (RMSE) metrics and root mean square error (MAE) metrics for single-step prediction reaching 46.08 and 26.22 at 60 min, respectively. In multi-step testing, the proposed model exhibits a 14.71 % reduction in RMSE error at the 15-min scale and a 9.29 % reduction at the 60-min scale compared to the benchmark model. This performance improvement enhances the operational efficiency and reliability of the energy storage plant, particularly under dynamic changes in the time series.

Multi-energy CT material decomposition using graph model improved CNN

In spectral CT imaging, the coefficient image of the basis material obtained by the material decomposition technique can estimate the tissue composition, and its accuracy directly affects the disease diagnosis. Although the precision of material decomposition is increased by employing convolutional neural networks (CNN), extracting the non-local features from the CT image is restricted using the traditional CNN convolution operator. A graph model built by multi-scale non-local self-similar patterns is introduced into multi-material decomposition (MMD). We proposed a novel MMD method based on graph edge-conditioned convolution U-net (GECCU-net) to enhance material image quality. The GECCU-net focuses on developing a multi-scale encoder. At the network coding stage, three paths are applied to capture comprehensive image features. The local and non-local feature aggregation (LNFA) blocks are designed to integrate the local and non-local features from different paths. The graph edge-conditioned convolution based on non-Euclidean space excavates the non-local features. A hybrid loss function is defined to accommodate multi-scale input images and avoid over-smoothing of results. The proposed network is compared quantitatively with base CNN models on the simulated and real datasets. The material images generated by GECCU-net have less noise and artifacts while retaining more information on tissue. The Structural SIMilarity (SSIM) of the obtained abdomen and chest water maps reaches 0.9976 and 0.9990, respectively, and the RMSE reduces to 0.1218 and 0.4903 g/cm3. The proposed method can improve MMD performance and has potential applications.

MCNMF-Unet: a mixture Conv-MLP network with multi-scale features fusion Unet for medical image segmentation

Recently, the medical image segmentation scheme combining Vision Transformer (ViT) and multilayer perceptron (MLP) has been widely used. However, one of its disadvantages is that the feature fusion ability of different levels is weak and lacks flexible localization information. To reduce the semantic gap between the encoding and decoding stages, we propose a mixture conv-MLP network with multi-scale features fusion Unet (MCNMF-Unet) for medical image segmentation. MCNMF-Unet is a U-shaped network based on convolution and MLP, which not only inherits the advantages of convolutional in extracting underlying features and visual structures, but also utilizes MLP to fuse local and global information of each layer of the network. MCNMF-Unet performs multi-layer fusion and multi-scale feature map skip connections in each network stage so that all the feature information can be fully utilized and the gradient disappearance problem can be alleviated. Additionally, MCNMF-Unet incorporates a multi-axis and multi-windows MLP module. This module is fully end-to-end and eliminates the need to consider the negative impact of image cropping. It not only fuses information from multiple dimensions and receptive fields but also reduces the number of parameters and computational complexity. We evaluated the proposed model on BUSI, ISIC2018 and CVC-ClinicDB datasets. The experimental results show that the performance of our proposed model is superior to most existing networks, with an IoU of 84.04% and a F1-score of 91.18%.

Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea

Scoring polysomnography for obstructive sleep apnea diagnosis is a laborious, long, and costly process. Machine learning approaches, such as deep neural networks, can reduce scoring time and costs. However, most methods require prior filtering and preprocessing of the raw signal. Our work presents a novel method for diagnosing obstructive sleep apnea using a transformer neural network with learnable positional encoding, which outperforms existing state-of-the-art solutions. This approach has the potential to improve the diagnostic performance of oximetry for obstructive sleep apnea and reduce the time and costs associated with traditional polysomnography. Contrary to existing approaches, our approach performs annotations at one-second granularity. Allowing physicians to interpret the model's outcome. In addition, we tested different positional encoding designs as the first layer of the model, and the best results were achieved using a learnable positional encoding based on an autoencoder with structural novelty. In addition, we tried different temporal resolutions with various granularity levels from 1 to 360 s. All experiments were carried out on an independent test set from the public OSASUD dataset and showed that our approach outperforms current state-of-the-art solutions with a satisfactory AUC of 0.89, accuracy of 0.80, and F1-score of 0.79.

HGT: A Hierarchical GCN-Based Transformer for Multimodal Periprosthetic Joint Infection Diagnosis Using Computed Tomography Images and Text

Prosthetic joint infection (PJI) is a prevalent and severe complication characterized by high diagnostic challenges. Currently, a unified diagnostic standard incorporating both computed tomography (CT) images and numerical text data for PJI remains unestablished, owing to the substantial noise in CT images and the disparity in data volume between CT images and text data. This study introduces a diagnostic method, HGT, based on deep learning and multimodal techniques. It effectively merges features from CT scan images and patients' numerical text data via a Unidirectional Selective Attention (USA) mechanism and a graph convolutional network (GCN)-based Feature Fusion network. We evaluated the proposed method on a custom-built multimodal PJI dataset, assessing its performance through ablation experiments and interpretability evaluations. Our method achieved an accuracy (ACC) of 91.4% and an area under the curve (AUC) of 95.9%, outperforming recent multimodal approaches by 2.9% in ACC and 2.2% in AUC, with a parameter count of only 68 M. Notably, the interpretability results highlighted our model's strong focus and localization capabilities at lesion sites. This proposed method could provide clinicians with additional diagnostic tools to enhance accuracy and efficiency in clinical practice.

HyFormer: Hybrid Transformer and CNN for Pixel-Level Multispectral Image Land Cover Classification

To effectively solve the problems that most convolutional neural networks cannot be applied to the pixelwise input in remote sensing (RS) classification and cannot adequately represent the spectral sequence information, we propose a new multispectral RS image classification framework called HyFormer based on Transformer. First, a network framework combining a fully connected layer (FC) and convolutional neural network (CNN) is designed, and the 1D pixelwise spectral sequences obtained from the fully connected layers are reshaped into a 3D spectral feature matrix for the input of CNN, which enhances the dimensionality of the features through FC as well as increasing the feature expressiveness, and can solve the problem that 2D CNN cannot achieve pixel-level classification. Secondly, the features of the three levels of CNN are extracted and combined with the linearly transformed spectral information to enhance the information expression capability, and also used as the input of the transformer encoder to improve the features of CNN using the powerful global modelling capability of the Transformer, and finally the skip connection of the adjacent encoders to enhance the fusion between different levels of information. The pixel classification results are obtained by MLP Head. In this paper, we mainly focus on the feature distribution in the eastern part of Changxing County and the central part of Nanxun District, Zhejiang Province, and conduct experiments based on Sentinel-2 multispectral RS images. The experimental results show that the overall accuracy of HyFormer for the study area classification in Changxing County is 95.37% and that of Transformer (ViT) is 94.15%. The experimental results show that the overall accuracy of HyFormer for the study area classification in Nanxun District is 95.4% and that of Transformer (ViT) is 94.69%, and the performance of HyFormer on the Sentinel-2 dataset is better than that of the Transformer.

Comprehensive Evaluation of Government Economic Management Performance Based on Multidimensional Data Mining in Fuzzy Comprehensive Environment

Economic, political, social, and military activities all fall under the umbrella of government duties. The modification and reinterpretation of economic functions are the primary focus of the innovation in government administration style under the wave of economic globalisation. The effectiveness of the government's economic administration determines the general level of public administration at the federal level. An essential issue of national public administration that has a significant impact on the political growth of many nations is the performance evaluation of government economic administration. The people are the subject of government economic administration performance evaluation, and the people adjudicate the outcomes. An essential method of assessing the productivity of various departments is through performance review. The nation is currently in the process of switching from its long-standing planned economic system to a market economic system. The internal and external environments of governmental organisations are continually changing in addition to the strong trend of economic globalisation. Determining and advancing the national government's economic administration model is so crucial. The article offers a multidimensional data mining-based optimum design scheme for the thorough assessment of government economic administration performance. The fuzzy comprehensive evaluation rule is an effective method for quantifying the qualitative indicators when the quantitative indicators in the evaluation index system are difficult to measure. It is relatively simple, reasonable, and simple to operate in practise, which is conducive to the thorough and scientific performance evaluation of the government economic administration's science and technology administration functions. Following an assessment of the system's performance using association rule data mining technologies, a simulation test analysis is completed. The accuracy of the proposed arithmetic, which is 8.26% higher than the conventional arithmetic, is demonstrated by simulation results. The development of an evaluation model that incorporates both subjective and objective criteria, as well as the thorough assessment of the effectiveness of government economic administration based on data mining technologies, has excellent application prospects and practical value.

A Comprehensive Assessment of Cultivation Environment of Top Innovative High-Level Talents Based on Deep Learning Algorithm

The quality of talent has increased across all fields due to the constant growth of different industries and the growing job saturation. Real-time job information on recruitment platforms can, therefore, accurately reflect the demand for talent from businesses, serving as a basis for the creation of training policies in schools. In international competition, the development of talents, especially top-level talents, will become more and more crucial. Growing in importance is China's economy and social development. The evaluation of higher vocational and technical talents, however, should also be assessed from a variety of angles, given the diversification of talent training objectives and teaching methods, as well as the expansion of teaching functions. An emerging machine learning technology called deep learning (DL) has been developed to bring machine learning closer to the goals of artificial intelligence. This essay offers a thorough evaluation of the depth of deep learning as it relates to the development of innovative talent in schools. The entire school must be strengthened. It is demonstrated that the average execution time is slashed by 0.0024 s, and the learning sample size error of the DL model is reduced by 0.05276 when compared to the Apriori method. As a result, implementing and researching the DL model can significantly improve both the overall teaching quality of schools and their capacity for innovation.

Construction of a Basic Japanese Teaching Resource Base Based on a Deep Neural Network under a Big Data Environment

A challenge for education and teaching in universities is posed by “Internet plus,” which has made numerous educational resources at universities richer and more accessible. The development of a professional Japanese teaching resource base should be centered on the needs and characteristics of Japanese teaching in universities, as well as establish and enhance the mechanism for resource base construction. All forms of instructional resources should also continuously be updated and improved in order to realize the diversified, systematic, open, and long-term development of Japanese instructional resources. In light of the current state of the information technology industry’s rapid expansion, this essay examines a few issues with the building of a Japanese teaching resource database. A fundamental Japanese teaching resource database built on DNN was created as a result. The CNN technology is used in this study to create the Arduino device identification application. Utilizing gadgets in the learning process, learners can obtain learning resources using the Arduino device identification program before engaging in learning activities. The experimental findings also demonstrate that the precision rate and recall rate of the Japanese teaching resource database system developed in this study may achieve about 93 and 94 percent, respectively. Its performance is better than the conventional teaching resource system, and it can offer top-notch teaching resources for teaching fundamental Japanese.

Construction of College English Teaching Environment Assessment Model Based on BP Neural Network and Multiple Intelligence Theory

College English has almost always been a required course, and a college student's level of English proficiency is one of the factors used to assess their learning capacity. The quality of students' English learning is largely influenced by the level of English instruction provided in colleges. However, there are still a lot of issues with college English instruction today, the most glaring of which is that English instruction is being overly simplified, and that the methods, modes, and purposes of instruction are also very narrow. Due to this, it is challenging for most colleges and universities' English teaching levels to satisfy the requirements of high-level education. The PSO-BP neural network model, which optimizes the BP neural network (BPNN), is used in this study to build a high-precision and diversified English teaching evaluation model in order to address the aforementioned issues. According to the experimental findings, the PSO-BPNN algorithm has a relative error of just 0.29 percent and an average accuracy rate of 97.02 percent. Overall performance is superior to that of the conventional BPNN algorithm, and it is the most adaptable in terms of creating various evaluation modes.

Cross-Cultural Education and College English Speculative Reading Teaching in Multi-Modal Theoretical Environment

Due to the extensive development of globalization, the ethnic makeup of the world is becoming more and more complex, and the issue of cultural diversity has emerged as a key concern for all nations' educational systems. Cross-cultural education is an educational prescription for many western countries to deal with cultural diversity and strengthen national cohesion in order to address this issue and promote mutual respect, understanding, and communication between different groups and individuals. One of the most recent developments in global education is cross-cultural education, which has also given rise to a brand-new area of study in the field of education. The research has identified cross-cultural education as making up about 50% of all education. Numerous nations have practiced cross-cultural education, which has been promoted by UNESCO, with many issues that merit study. The traditional English teaching approach cannot meet the learning needs of today's students in light of the recent curriculum reform, and students have long engaged in passive learning. College English instructors must modify traditional teaching philosophies, adapt to contemporary developments, integrate information technology, use a variety of teaching techniques, fully pique students' interest in learning, and encourage them to take the initiative in their own education if they want to reclaim the initiative from the students in their classrooms.

Tourism Demand Prediction Model Using Particle Swarm Algorithm and Neural Network in Big Data Environment

Since demand forecasting is the first step in managing and operating a tourism business, its accuracy is very important to tourism businesses. In order to address NN's drawbacks, such as local optimization, slow convergence, and large sample sizes, this paper organically combines the PSO and NN models and builds a PSO-NN-based tourism demand forecasting model. The tourism demand forecasting indexes, the choice of NN forecasting models, the modelling process, and the implementation methods are first analysed and studied along with the fundamental theories and forecasting techniques of PSO and NN. In order to increase the precision of the prediction model, the PSO algorithm is also used to optimise the weights and thresholds of the NN. The final section of the paper compares the performance of the model developed in this paper with the most widely used model for forecasting tourism demand. According to the experimental findings, this model's prediction accuracy can reach 95.81 percent, or about 10.09 percent higher than the prediction accuracy of the conventional NN model. There are some practical implications to this research. Applying the optimization model to the forecast of tourism demand is doable and practical.

Evaluation System of Music Art Instructional Quality Based on Convolutional Neural Networks and Big Data Analysis

In order to speed up the process of high-quality education and improve the level of education quality among the general public, people have pushed for the use of music art education in recent years. In this respect, this study covers the CNN-based assessment of the quality of music art teaching and creates a set of evaluation indices for that quality. The model architecture, network topology, learning parameters, and learning algorithm are all determined using this information, which also acts as the basis for the NN assessment model. The MATLAB simulation tool uses the CNN assessment model to train and learn a predetermined quantity of instructional quality data. The training experiment shows that this system can outperform other comparative systems in prediction accuracy by roughly 95%. Additionally, both the training and prediction accuracy of the model are completely acceptable. The evaluation findings and analytical data of the music art instructional quality assessment system created in this study can be used as a guide for determining the music art instructional quality and for making judgments regarding it.

A Novel English Translation Model in Complex Environments Using Two-Stream Convolutional Neural Networks

Although translation is an essential component of learning English, it does not receive the attention it merits in the modern English classroom. Teachers and students primarily emphasize listening, reading, and writing while neglecting the development of translation skills. The English test in China now reflects the fact that there are now very specific requirements for students' translation skills. As a result, we should emphasize developing students' translation skills when teaching them English. The following experimental data can be obtained following the study and experiment on the English translation simulation model based on the two-stream convolutional neural network: English vocabulary and grammar have passing and excellent rates of 90 and 57 percent, respectively, while reading has passing and excellent rates of 69 and 8 percent, respectively. The ability of students to translate into English has significantly improved after using the English translation simulation model based on the two-stream convolutional neural network.

Dance-Specific Action Recognition Method Based on Double-Stream CNN in Complex Environment

Technology for dance-specific motion recognition is widely used in many industries, but Chinese research in this area is still in its early stages. Recognizing specific dance movements is the key to learning about and comprehending human actions and behaviors. The fault-tolerant feature of standardized sign language recognition is extended under the condition of small sample sizes, but the recognition accuracy remains a challenge. This issue needs to be resolved by fusing the essential details of particular dance movements. A dual-stream convolution neural network is suggested in this paper to investigate the recognition of particular dance movements. In this paper, a dual-stream convolution neural network is used to study the recognition of particular dance movements. The time spent by this algorithm gradually increases as the number of people in the image does, but only slightly. The algorithms proposed by Bergonzoni (2017) and Liu et al. (2021) both experience linear increases in running time as the population grows. In contrast, the running time of the algorithm in this study essentially increases negligibly. It has become a problem deserving in-depth study. Double-stream convolution neural network improves the practical value and technical complexity of dance motion automatic generation technology in art and cultural heritage protection, dance teaching, dance video retrieval, and dance arrangement.

Big Data Analysis of Benign Interaction of Great Power Relations and New International Relations Based on Deep Learning

The development of a new type of international relations is the advancement and improvement of diplomatic thinking among contemporary nations. It also serves as a crucial yardstick for assessing the future global pattern and the direction of order changes. Proper interaction between major powers can foster the growth of new international relations and has a significant impact on advancing global cooperation and the promotion of human peace. The goal of this essay is to examine how friendly interactions between major powers have affected the development of new international relations. A deep learning network model is presented for this purpose. The deep learning model was used to identify the emotions of the survey results, analyze each person's emotional tendencies, and summarize and compare the data. Relevant questionnaire surveys were conducted using the online questionnaire survey method on individuals in various countries. The survey results in this paper demonstrate that 96.5 percent of Chinese, 89.3 percent of Russians, and 81.6 percent of Americans support friendly relations between major nations. Only a very small percentage of the investigators supported hostile relations, with their support being 1.06 percent, 3.11 percent, and 2.94 percent, respectively. Therefore, creating a win-win partnership between major powers is exactly what the people of all nations are calling for. In contrast to the past, it is no longer hostile and violent. People anticipate that more great powers will coexist peacefully.

The Collection and Utilization of Web Resources for Teaching World History Based on Data Mining Technology

The foundation of research for academics involved in world history education and research is timely access to pertinent foreign information, comprehension of domestic and international academic developments, and access to fundamental historical materials and research outputs from our ancestors. As a result, it is essential to investigate the necessity and viability of using network resources on the basis of analysis in order to respond to the educational philosophy of the new curriculum reform and adapt to the development of modern teaching methods based on network technology. The creation and use of curriculum resources is an essential component of curriculum development and a crucial assurance for curriculum implementation. The growth and acceptance of online learning will undoubtedly influence how history education is practised around the world. The scarcity of online learning resources is currently a bottleneck impeding the growth of online education. On the other hand, data mining (DM) takes the massive amount of incomplete data and extracts the useful knowledge and information hidden within it. This paper explores the “DM” process of utilizing online resources and suggests a method for gathering and utilizing world history education online resources based on DM technology. The experimental results show that the test interval between MapReduce and DM gradually increases with the increase of data volume. The advantage of DM is more obvious, as the average test time of DM is 27.66 seconds shorter than that of MapReduce. Therefore, DM has high application value in the field of search engines and social network analysis.

Image Recognition of Sports Athletes’ High-Intensity Sports Injuries Based on Binocular Stereo Vision

Sports athletes are prone to certain injuries during high-intensity exercise training. In the process of treating an injury, images of the injury site need to be collected and identified. However, the traditional recognition method cannot effectively extract the features of the image. At the same time, it ignores the optimization of the damage image recognition results, resulting in low recognition accuracy and poor efficiency. Binocular stereo vision technology can quickly and accurately detect moving objects. Therefore, in order to more accurately identify high-intensity sports injury images, this study takes the high-intensity sports injury images as the basic research object. Several processes of image processing based on binocular stereo vision are analyzed, and the vulnerable parts of the body in high-intensity sports are also studied. Finally, the method in this study is verified. The experimental results show that the method proposed in this study reduces the average error rate by 0.19% compared with the traditional recognition method. It can effectively identify and detect injury images, thereby improving the accuracy and stability of sports injury image identification. The  identification time is also shortened accordingly, which has certain practicability and feasibility. In addition, the binocular camera used in this study has high accuracy, and the obtained images of sports injuries are of good quality, which lays a foundation for image detection and recognition.

Construction of Digital Platform of Religious and Cultural Resources Using Deep Learning and Its Big Data Analysis

This article analyzes the difficulties associated with the preservation and transmission of religious cultural resources and the difficulties encountered in the new development environment and background. It does so in light of the current state of religious, cultural resources. The protection, growth, and use of religious and cultural resources against the backdrop of the digital era are elaborated upon and critically analyzed in this article. Based on the foregoing discussion, this article conducts a thorough analysis of the development of a digital platform for religious and cultural resources and its big data analysis, and it also suggests an image feature extraction algorithm based on DL. This article develops a clustering CNN based on the network with PCA vector as convolution kernel, which clusters small images and computes principal component vectors according to categories, generating multiple groups of convolution kernels to extract more features so that the input image can select feature extractors adaptively. Simulation and comparative analysis are used in this article to confirm the algorithm's effectiveness. Compared to the conventional NN algorithm, simulation results indicate that this algorithm is more accurate, with a maximum accuracy of about 95.14 percent. It has some reference value for the research that will be done in relation to the creation of the next digital platform for religious and cultural resources.

Optimization Algorithm of Urban Rail Transit Network Route Planning Using Deep Learning Technology

Under the present background, optimizing the existing urban rail transit network is the focus of urban rail transit construction at present. Based on DL, this paper constructs the optimization algorithm of urban rail transit network route planning. According to the current urban layout and urban planning, build a suitable rail transit network line form; according to the function, the types of urban rail transit stations are divided, and the optimization of urban rail transit network lines is realized. In addition, according to the K short path algorithm, this paper calculates the effective path between any stations of rail transit and, according to the model, allocates the passenger flow to each path. Experimental results show that the accuracy of real-time traffic flow prediction by this algorithm can reach 94.98%, which is about 9% higher than other methods. This algorithm can effectively optimize the route planning of urban rail transit network. This verifies the effectiveness of the route planning optimization algorithm proposed in this paper. Using the algorithm in this paper for line planning can get good real time, rationality, and optimality.

Reset and Integration of Music Instructional Resources Using Deep Convolutional Neural Networks

In order to overcome the problem that learners and teachers cannot find instructional resources to meet their needs and information overload in the massive resources, this article proposes and designs a music instructional resource management platform based on DCNN. This article expounds the overall goal, design principle, overall structure, and interface design of the system. At the same time, the whole construction process of a music instructional resources integration system based on DCNN is discussed in detail from the aspects of configuration of development environment, localization of platform interface, and realization of main functions of the system. In addition, through the demand analysis tool, the demand of college music instructional resources management is analyzed in detail and deeply, and the demand document is formed. This article makes an in-depth study on the categories of music instructional resources and summarizes the resource classification methods that are in line with the actual instructional activities. The experiments show that the accuracy of the proposed algorithm is improved by about 6% compared with the fuzzy clustering algorithm. At the same time, the stability of this system can reach 96.14%. This system is rich in functions and easy to use and can provide a feasible scheme for the management of instructional resources in various disciplines.

A Novel Piano Arrangement Timbre Intelligent Recognition System Using Multilabel Classification Technology and KNN Algorithm

In this paper, melody and harmony are regarded as the task of machine learning, and a piano arranger timbre recognition system based on AI (Artificial Intelligence) is constructed by training a series of samples. The short-time Fourier transform spectrum analysis method is used to extract the piano timbre characteristic matrix, and the electronic synthesis of timbre recognition is improved by extracting the envelope function. Using the traditional multilabel classification method and KNN (K-nearest neighbor) algorithm, a combined algorithm of these two algorithms is proposed. The experimental results show that the detection rate increases from 61.3% to 70.2% after using the combined classification algorithm. The correct rate also increased from 40.3% to 48.9%, and the detection rate increased to 74.6% when the K value was set to 6. The experimental results show that, compared with the traditional classification algorithm, this algorithm has a certain improvement in recognition rate. Using this system to recognize the timbre of piano arrangement has a high recognition accuracy, which is worthy of further popularization and application.

Model of Markov-Based Piano Note Recognition Algorithm and Piano Teaching Model Construction

Piano note recognition is a process that converts music audio files into digital music files automatically, which is critical for piano assistant training and automatic recording of musical pieces. The Merle spectral coefficients, for example, have been used to implement the majority of the existing examples. The piano is one of the most popular forms of student education in today’s world. Piano teachers should be aware of the implications. We can only truly adapt piano teaching to the educational purposes of higher education institutions if we implement a systematic, progressive, practical, and innovative philosophy of piano teaching. The Markov model is a statistical model that is widely used in speech signal processing. This thesis develops a set of mathematical models for piano speech recognition based on the Markov model, learns them systematically and scientifically, and achieves a better teaching effect. It is demonstrated that the Markov method detects the corresponding endpoints with an accuracy of 72.83 percent, which is 16.42 percent better than the a priori method. In terms of amplitude and phase, the Markov model shows a significant improvement. The findings of this study can be used to improve piano playing techniques taught to students in accordance with their favourite popular music, depending on the theme.

Optimization Model of Mathematics Instructional Mode Based on Deep Learning Algorithm

This paper proposes corresponding teaching methods and instructional modes based on predecessors' research on mathematics instructional mode and the current state of mathematics teaching. In addition, this paper constructs a teaching evaluation model based on DL algorithm based on an in-depth study of DL-related theories in order to accurately and scientifically analyze the problems that exist in mathematics teaching. This paper constructs an instructional quality evaluation index system based on rationality and fairness, and uses the BPNN evaluation model to train and study a set of instructional quality data. Finally, the experimental results show that this system has a high level of stability, with a 96.37 percent stability rate and a 95.42 percent evaluation accuracy rate. The results of this paper's evaluation of the mathematical instructional quality model are objective and reasonable. It can accurately assess instructional quality while also assessing problems in the teaching process based on the instructional quality scores and making reasonable recommendations for teaching improvement based on the weak links in the teaching process. It has the potential to provide a workable system for assessing instructional quality.