检索结果-内蒙古大学图书馆

Cognitive Computing and Complex Data (ICCD), International Conference on the

作者： Hui Zong Yun Zhao Heng Zhou Le Zhang Faculty of Computer and Software Engineering Huaiyin Institute Of Technology Huaian China

Chemical accident news data encompasses essential information such as news headlines, news content, and news sources, with the context of news content playing a crucial role. To enhance the accuracy of text feature extraction and improve the efficiency of chemical accident news classification, this paper introduces a feature fusion-based classification approach. The proposed model employs a multi-layer convolutional neural network (CNN) to extract local features from the text of chemical accident news. Additionally, a Bidirectional Long Short-Term Memory (BiLSTM) network is utilized to capture global features, supplemented by the integration of a Self-Attention mechanism behind the BiLSTM network to assign weights to the features and reduce noise. The local and global features are then fused to enrich the semantic information. Furthermore, the feature fusion information undergoes maximum pooling and average pooling to reduce dimensionality and enhance the training speed. Finally, the information is fed into a Softmax layer for classification. Experimental results demonstrate that the proposed neural network model, namely ABLSACNN (Add-CNN-BiLSTM-Self-Attention), outperforms the CNN-Self-Attention model. The ABLSACNN model exhibits an improvement of1.59% in accuracy, 2.46% in recall rate, and 1.93% in F1 score on the chemical accident news dataset, thereby showcasing its superiority.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Semi-Anchoring Guided High-Resolution Capsule Network for Vehicle Logo Recognition

A Semi-Anchoring Guided High-Resolution Capsule Network for ...

引用

Cognitive Computing and Complex Data (ICCD), International Conference on the

作者： Lanfang Wang Zhenyang Yu Yongtao Yu Faculty of Computer and Software Engineering Huaiyin Institute of Technology Huaian China

As a unique identity reflecting the manufacturer of the vehicle, the vehicle logo information plays an important role in many transportation-related applications. However, due to the challenges of size variations, shape and form diversities, deformations, occlusions, and complex scenarios, it is still not an easy task to realize highly accurate vehicle logo recognition from images. This paper proposes a novel semi-anchoring guided high-resolution capsule network (SAGHR-CapsNet) for vehicle logo recognition. First, constructed with a multibranch high-resolution capsule network architecture functioned with repeated multiresolution feature fusion for feature extraction, the SAGHR-CapsNet can extract semantically strong and spatially accurate feature representations at each scale. Second, designed with a capsule-based efficient self-attention mechanism for feature semantic promotion, the SAGHRCapsNet functions excellently to attend to channel-wise informative features and target-oriented spatial features. Finally, adopted with a semi-anchoring guided strategy for vehicle logo recognition, the SAGHR-CapsNet performs promisingly to simultaneously improve the processing efficiency and guarantee the recognition accuracy. Intensive quantitative evaluations and comparative analyses on two large-scale data sets demonstrated the applicability and superiority of the SAGHR-CapsNet in vehicle logo recognition tasks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Air Quality Index Prediction Based on CNN-LSTM-Attention Hybrid Modeling

Air Quality Index Prediction Based on CNN-LSTM-Attention Hyb...

引用

Cognitive Computing and Complex Data (ICCD), International Conference on the

作者： Yinyin Li Tao Jiang Hanzhu Gu Weibing Lu Qingguo Wu Yongtao Yu Faculty of Computer and Software Engineering Huaiyin Institute of Technology Huaian China

Air Quality Index (AQI) is an important indicator for determining good or bad air quality. The accurate and efficient prediction of AQI plays a positive role in promoting the management of air pollution. However, current algorithms for predicting AQI usually do not comprehensively consider the effects of pollutant factors and meteorological factors on the prediction performance. Therefore, taking pollutant factors and meteorological factors as the basis of the model study, a CNNLSTM-Attention hybrid model is proposed. The CNN-LSTM module is used to obtain the air quality-related features, and the attention mechanism is introduced to weigh and sum the output of the LSTM in order to obtain the final attention-weighted features. The results show that the model has better performance than the single model for the prediction of the air quality index.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Ternary-attention Capsule Feature Pyramid Network for Building Footprint Extraction from Remote Sensing Imagery

A Ternary-attention Capsule Feature Pyramid Network for Buil...

引用

Cognitive Computing and Complex Data (ICCD), International Conference on the

作者： Haiyan Zhang Yazhou Li Meifeng Wang Faculty of Computer and Software Engineering Huaiyin Institute of Technology Huaian China

The up-to-date and accurate building footprint database plays a significant role in a large variety of applications. Recently, remote sensing images have provided an important data source for building footprint extraction tasks. However, due to topology variations, color diversities, and complicated rooftop and environmental scenarios, it is still a challenging task to realize fully automated and highly accurate extraction of building footprints from remote sensing images. In this paper, we propose a novel ternary-attention capsule feature pyramid network (TA-CapsFPN), which is formulated with a capsule feature pyramid network architecture and integrated with context-augmentation and feature attention modules, aiming at improving the building footprint extraction accuracy by combining the superior properties of capsule representations and the powerful capability of attention mechanisms. Quantitative evaluations and comparative analyses show that the TA-CapsFPN provides a promising and competitive performance in processing buildings of varying conditions.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Application of Lightweight DSC-NET Network Model in Histopathological Images of Lung Cancer and Colon Cancer

Application of Lightweight DSC-NET Network Model in Histopat...

引用

Cognitive Computing and Complex Data (ICCD), International Conference on the

作者： Wenlong An Hui Zong Heng Zhou Jian Wang Faculty of Computer and Software Engineering Huaiyin Institute Of Technology Huaian China

Classification of histopathological images is a fundamental task in the workflow of pathological diagnosis. Due to the complexity of pathological images, it is particularly important to use deep learning to improve diagnostic efficiency. This paper designs a lightweight network model DSC-NET, which consists of multi-scale feature stitching Multi-Conv, coordinate attention CA. The improved selfcalibrated convolution MSC-Conv is composed of selfcalibrated convolution SC-Conv, coordinate attention CA and depthwise DW. In addition, the DSC-NET network model converts the 1×1 convolutional layer (Conv) in the Block module into a linear layer (Linear), which reduces the computational complexity of the model while maintaining the ability of the convolution operation to capture local features. The research in this paper adopts the lung cancer and colon cancer datasets and adds Gaussian noise to these datasets to simulate the equipment shooting situation and evaluate the lightweight DSC-NET network model. Through quantitative comparisons with previous state-of-the-art methods, our experimental results demonstrate that the proposed method achieves superior accuracy. Furthermore, our method stands out with a smaller parameter count and significantly lower FLOPs, highlighting its efficiency and computational advantages. It has important potential to assist pathologists in pathological diagnosis.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Surface Defect Detection Method Based on Attention Mechanism for Workpieces

A Surface Defect Detection Method Based on Attention Mechani...

引用

Cognitive Computing and Complex Data (ICCD), International Conference on the

作者： Junlin Qiu Yongshan Shen Wanli Feng Haodong Zhang Keke Sun Yuxin Qin Faculty of Computer and Software Engineering Huaiyin Institute of Technology Huai’an China

With the rapid development of image processing technology in recent years, faced with the issues of low detection accuracy and missed detections in the process of surface defect detection on workpieces, we propose a workpiece surface defect detection method based on attention mechanism. The model uses EfficientDet-d0 as the baseline network, mixes the Fused-MBConv structure and MBConv structure in EfficientNetv2 network as part of the feature extraction network, and uses the convolutional attention module CBAM to focus on the information of space and channel direction at the same time, and the Hardmish activation function is used in the structure. Introducing a fast spatial pyramid pooling module (SPPF) at the top of the feature extraction network increases the network’s depth and enhances its expressive power. The extracted features are fed into the Improved-BiFPN (Bidirectional Feature Pyramid Network) for feature enhancement, improving the model’s ability to detect defects of different sizes. Experimental results demonstrate that our proposed surface defect detection method for workpieces outperforms other advanced detectors.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Travelling Recommendation System based on Knowledge Graph 3

A Travelling Recommendation System based on Knowledge Graph

引用

3rd International Academic Exchange Conference on Science and technology Innovation, IAECST 2021

作者： Dai, G.R. Xu, J.Y. Zhou, H. Zhou, L. Cao, Z.H. Zhu, M.D. Huaiyin Institute of Technology Faculty of Computer and Software Engineering China

ISBN: (纸本)9781665402675

In recent years, the improvement of people's live standard lead to an increasing demand for travelling, but the information on scenic spots on the Internet is ponderous and the accuracy of scenic spot recommendations is not high. This study aims to solve these problems to realize a more accurate scenic spot recommendation for self-drivings and provide a better visualization of recommendation results. A recommendation model based on the Multi-Task Feature Learning for Knowledge Graph Enhanced Recommendation (MKR) algorithm is designed in this study. Then, based on this model, an intelligent knowledge graph based recommendation system for scenic spots is constructed. Rich unstructured text data of scenic spots in major scenic spot websites are crawled using Selenium and are stored and managed by the neo4j graph database. Through experimental simulation, the recommendation accuracy can reach over 84%. Compared to the Propagating User Preferences on the Knowledge Graph for Recommender Systems (RippleNet), the accuracy rate is enhanced by 6.2%. © 2021 IEEE.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Transaction transmission model for blockchain channels based on non-cooperative games

引用

Science China(Information Sciences) 2023年第1期66卷 97-114页

作者： Peiyun ZHANG Chenxi LI Mengchu ZHOU Wenjun HUANG Abdullah ABUSORRAH Omaimah O.BAMASAG Engineering Research Center of Digital Forensics of Ministry of Education School of Computer and SoftwareNanjing University of Information Science and Technology School of Computer and Information Anhui Normal University Department of Electrical and Computer Engineering New Jersey Institute of Technology Center of Research Excellence in Renewable Energy and Power Systems Department of Electrical and Computer Engineering Faculty of Engineering and K.A.CARE Energy Research and Innovation CenterKing Abdulaziz University Department of Computer Science Faculty of Computing and Information Technology Center of Excellence in Smart Environment Research King Abdulaziz University

In blockchain networks, transactions can be transmitted through channels. The existing transmission methods depend on their routing information. If a node randomly chooses a channel to transmit a transaction, the transmission may be aborted due to insufficient funds(also called balance) or a low transmission rate. To increase the success rate and reduce transmission delay across all transactions, this work proposes a transaction transmission model for blockchain channels based on non-cooperative game *** balance, channel states, and transmission probability are fully considered. This work then presents an optimized channel transaction transmission algorithm. First, channel balances are analyzed and suitable channels are selected if their balance is sufficient. Second, a Nash equilibrium point is found by using an iterative sub-gradient method and its related channels are then used to transmit transactions. The proposed method is compared with two state-of-the-art approaches: Silent Whispers and Speedy Murmurs. Experimental results show that the proposed method improves transmission success rate, reduces transmission delay,and effectively decreases transmission overhead in comparison with its two competitive peers.

关键词： Blockchain channel transaction transmission channel balance channel state channel transmission probability non-cooperative game

来源：评论

学校读者我要写书评

暂无评论

Safety Helmet Detection Algorithm for Complex Scenarios Based on PConv-YOLOv8

Safety Helmet Detection Algorithm for Complex Scenarios Base...

引用

Cognitive Computing and Complex Data (ICCD), International Conference on the

作者： Yuanyuan Wang Feilong Jiang Yazhou Li Haiyan Zhang Meifeng Wang Shaofeng Yan Faculty of Computer and Software Engineering Huaiyin Institute of Technology Huaian China

Detecting safety helmets in complex environments is challenging due to issues like occlusion and lighting variations. Addressing the issues of slow detection speed and low object detection accuracy in complex environments with the YOLOv8 model, this paper introduces a lightweight safety helmet detection model, called PConv-YOLOv8, that is suitable for real-time applications in complex environments. Our method incorporates the PConv (Partial Convolution) module into the YOLOv8 model, reducing the complexity of the feature extraction network while enhancing feature representation accuracy. It also incorporates SimAM attention to extract and enhance the most relevant features by evaluating their similarity. Additionally, it considers category imbalance and positional regression in the target detection task, enhancing the model’s performance in target category identification and positional localization. Moreover, we propose the Wise-Distribution Focal Loss function to improve bounding box selection accuracy and enhance the model’s robustness. This paper introduces the Wise-Distribution Focal Loss method, which enhances the performance of target category recognition and location localization by improving the accuracy of bounding box selection and increasing the robustness of the overall model. The experimental results demonstrate that the method proposed in this paper achieves a 125% improvement in detection speed and a 1.8% increase in mAP0.5 compared to the YOLOv8 model.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Lightweight Protective Clothing Detection Algorithm Based on Ghost Convolution and GSConv Convolution

Lightweight Protective Clothing Detection Algorithm Based on...

引用

Cognitive Computing and Complex Data (ICCD), International Conference on the

作者： Qianjun Guan Kun Yu Haoran Wang Qiangqiang Li Shihong Gao Faculty of Computer and Software Engineering Huaiyin Institute of Technology Huai’an Jiangsu China

GhostNet is proposed by Huawei Noah’s Ark Laboratory in CVPR2020, which can be used with the same accuracy, with less speed and less computation than SOTA method. This convolution is a low-cost operation used to decrease the amount of computation and related parameters. Ghost convolution can be used as an alternative to traditional convolution operations because they have the same accuracy. However, utilize significantly less computation and a much smaller parameter count. GSConv(Group-Shared Convolution) is a convolution operation based on group shared weights. This operation achieves efficient parallel convolution operations by grouping convolution kernels. In this way, the parameters used and calculation complexity of the model can be reduced to some extent. This makes the network more computationally efficient and smaller in model size while maintaining accuracy. The algorithm can detect the location of protective clothing quickly and efficiently, and its accuracy and efficiency have been verified. Because the algorithm adopts lightweight network structure, real-time detection can be realized on lowpower devices efficiently. The model architecture enhances the performance based on the YOLOv5 target detection algorithm by incorporating additional features. So that the model can maintain a high precision, which is crucial for the protective clothing detection algorithm.

关键词：

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：