卷 6 ， 期 23 ， 2023
- 摘要：Despite pedestrian detection technology has become mature, dynamic image detection is still difficult, especially in the dynamic image, which needs fast image positioning and recognition. Aiming at the difficulty of dynamic image location and recognition, this paper proposed a fast location and recognition method based on YOLOv7. This method is different from other real-time target detection methods, which mainly focus on the efficient architecture design. YOLOv7 focuses on the optimization of the training process, and improves the accuracy of target location and recognition by introducing the bag-of-freebies module training method, which increases the training cost but does not increase the reasoning cost. The experiment shows that the method can effectively reduce the parameters of the real-time target detector by about 40% and the calculation amount by 50%, and has better positioning and recognition effect for dynamic images.关键词：YOLOv7;dynamic image;target location and identification;bag-of-freebies121|97|0更新时间：2025-03-26
- 摘要：The aim of this study is to construct a machine learning model for predicting the risk of aneurysm rupture by integrating hemodynamic and morphological features of patients with intracranial aneurysms. Cases of patients with intracranial aneurysms who sought treatment at the Department of Neurosurgery, Affiliated Hongqi Hospital of Mudanjiang Medical University from February 2021 to December 2023 were collected. The study group comprised 130 patients with ruptured aneurysms, while the control group included 60 patients with unruptured aneurysms. All patients provided CT (Computed Tomography) images in DICOM format, and after obtaining approval from the hospital’s ethics committee and signed informed consent from the patients’ families, three-dimensional reconstruction of the aneurysm DICOM format CT images was performed using MIMICS software. Morphological parameters were obtained through morphological measurements, and hemodynamic parameters were calculated through computational fluid dynamics simulations. The dataset for constructing the machine learning clinical prediction model was built by combining morphological and hemodynamic parameters. Ten machine learning algorithms were employed to construct the prediction models, and their performance was evaluated using metrics such as accuracy, AUC (Area Under the Curve), recall rate, and F1 score. Additionally, model interpretability was analyzed using the SHAP method. Among all the models tested, the random forest model demonstrated the best performance with an accuracy of 0.78, an AUC value of 0.81, and a high recall rate of 0.72. The machine learning model constructed by integrating the morphological and hemodynamic features of aneurysms can provide a powerful tool for clinical decision-making and shows promising potential for clinical application.关键词：Aneurysm;Rupture Risk;Hemodynamic Features;Morphological Features;machine learning;Model interpretability183|319|0更新时间：2025-03-26
- 摘要：Analyzing the characteristics of channels in rocks is important for assessing reservoir quality. Traditional methods for segmenting channels in rock thin section images mainly rely on binary images of pores for channel extraction. However, these methods tend to produce over-segmentation and low segmentation accuracy when analyzing complex pore structures, leading to poor generalization performance of the algorithm. To address these issues, this paper proposes a deep learning-based channel segmentation algorithm. The algorithm uses UNet++ as the backbone network for feature extraction. It incorporates CBAM (Convolutional Block Attention Module) attention modules during downsampling to enhance the model's ability to capture correlations between features of different dimensions and reduce semantic differences in skip connections. This improves the accuracy of channel feature extraction. Experimental comparisons of the proposed algorithm and other existing channel segmentation algorithms are conducted based on four metrics: CR (Correctness Rate), UR (Under-segmentation Rate), OR (Over-segmentation Rate), and MIoU (Mean Intersection over Union). The results show that the proposed method can effectively suppress channel over-segmentation and achieve higher accuracy in channel segmentation.关键词：cast thin section images;throat segmentation;UNet++;CBAM;skip connections54|132|0更新时间：2025-03-26
- 摘要：5G networks have the characteristics of low latency, high reliability, and wide coverage, making them the optimal choice to support industrial internet wireless networks. The integration of 5G into the industrial internet will significantly expand and enhance the capabilities of the industrial internet. The development momentum of China's “5G+Industrial Internet” is effective, and it is entering a new stage of industrial deepening and empowering development. However, there are still many problems in technology research, industry application, commercial profit models, network security protection. This article analyzes typical application cases of “5G+Industrial Internet” in different manufacturing industries in China, analyzes the innovative application development trend of “5G+Industrial Internet” empowering the manufacturing industry, and proposes policy recommendations for high-quality promotion of “5G+Industrial Internet” empowering the digital transformation of the manufacturing industry in China from the aspects of technology research and development, application expansion, network construction, security prevention, etc.关键词：5G;industrial internet;manufacturing industry;Digital Transformation439|269|0更新时间：2025-03-26
RESEARCH PAPER
- 摘要：This article designs a cloud based operation and maintenance big data management platform, which utilizes distributed computing and big data technology to achieve deep analysis and intelligent decision-making of operation and maintenance data, optimize operation and maintenance management processes, and improve operation and maintenance efficiency and quality. This platform has functions such as real-time monitoring, intelligent analysis, fault prediction, and spare parts management, providing automated, intelligent, and visual operation and maintenance services for enterprises and institutions, helping them achieve refined operation and maintenance and improve service quality.关键词：IT operation and maintenance management;Intelligent operation and maintenance;Cloud computing;Big data technology171|181|0更新时间：2025-03-26
- 摘要：In order to defend the ideological position, provide high-quality public services and promote the healthy development of the audio-visual industry, it is urgent to build an intelligent supervision platform for audio-visual programs. Research new changes in audio-visual program content creation, public opinion ecology, media structure, and communication methods under the background of media convergence, and propose to use local open source large language models combined with local audio-visual program supervision knowledge bases and applications to improve data collection, processing, storage, and application efficiency, solve the problems of many audio-visual program collection objects, diverse page layouts, little support for content research and judgment, long report writing time, and long equipment maintenance cycles, and explore new models for audio-visual program supervision in the era of media convergence.关键词：Large Language Model;audio-visual programs;smart supervision;content analysis;local knowledge base22|121|0更新时间：2025-03-26
CORRESPONDENCE
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09064830号-19
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.
- Total Visits：68219 Visits Today：1
0