检索结果-内蒙古大学图书馆

Environment Independent Gait Recognition Based on Wi-Fi Signals

IEEE Transactions on Mobile Computing 2025年第07期24卷 5934-5950页

作者： Yang, Wei Li, Zhixiang Chen, Sheng University of Science and Technology of China School of Computer Science and Technology Hefei230026 China University of Science and Technology of China Suzhou Institute for Advanced Research Suzhou215123 China Zhejiang Lab China

Gait recognition plays a pivotal role in the area of mobile computing. While various research approaches leverage images, radar, RF signals, pressure sensors, wearables, and other methods, utilizing Wi-Fi signals for gait recognition offers distinct advantages such as a wide sensing range, simple deployment, and passive sensing capabilities. However, traditional gait recognition systems relying on Wi-Fi signals often suffer from performance degradation due to variations in walking directions and environmental conditions. To address this issue, in this paper we propose EIGait, a gait recognition system based on Wi-Fi signal time-frequency spectrograms. EIGait enhances the robustness and generalizability of extracted features through spectrogram augmentation, self-contrastive learning, and domain-adversarial training. Particularly, improvements to ResNet in EIGait yield a Spectrogram ResNet, which is better suited for time-frequency spectrograms. In addition, using merely a single pair of Wi-Fi transmitter and receiver, and by minimal signal denoising, we achieve the state-of-the-art performance. To evaluate the performance of EIGait, we conduct extensive experiments. In a typical indoor environment, EIGait achieves F1 scores ranging from 98.11% to 98.31% for four to eight individuals. In cross-direction gait recognition, we obtain F1 scores of 96.64% to 94.45% for four to eight individuals. Moreover, under the more challenging conditions of cross-room gait recognition, EIGait attains F1 scores of 92.09% to 89.61% for four to eight individuals. Additionally, we conduct experiments on the public dataset 3.0, and the results also demonstrate significant superiority. © 2002-2012 IEEE.

关键词： Spectrographs

来源：评论

学校读者我要写书评

暂无评论

Properties of Ultra-low Thermal Expansion LAS Transparent Glass-ceramics Prepared by Spodumene

引用

Journal of Wuhan University of technology(Materials science) 2024年第3期39卷 541-550页

作者：何峰何子君 ZHOU Zhiqiang TIAN Yingliang ZHAO Zhiyong School of Materials Science and Engineering Beijing University of TechnologyBeijing 100124China Department of Functional Material Research Central Iron and Steel Research InstituteBeijing 100081China School of Materials Science and Engineering Wuhan University of TechnologyWuhan 430070China National Engineering Research Center of FPD Glass Technology Xianyang 712000China

The glass-ceramics were prepared with the spodumene mineral as the main raw material,and the effects of ZrO_(2)replacing TiO_(2)on the samples were systematically *** results show that the substitution of ZrO_(2)for TiO_(2)is not conductive to precipitate𝛽β-quartz solid solution phase,but can improve the transparency and flexural strength of *** the glass-ceramic with the highest visible light transmittance(87%)and flexural strength(231.80 MPa)exhibits an ultra-low thermal expansion of-0.028×10^(-7)K^(-1)in the region of 30-700℃.

关键词： ultra-low thermal expansion LAS transparent glass-ceramics substitution of ZrO_(2)for TiO_(2) spodumene mineral

来源：评论

学校读者我要写书评

暂无评论

Compete to Win: Enhancing Pseudo Labels for Barely-Supervised Medical Image Segmentation

引用

IEEE Transactions on Medical Imaging 2023年第11期42卷 3244-3255页

作者： Wu, Huimin Li, Xiaomeng Lin, Yiqun Cheng, Kwang-Ting The Hong Kong University of Science and Technology Department of Computer Science and Engineering Hong Kong Hong Kong The Hong Kong University of Science and Technology Department of Electronic and Computer Engineering Hong Kong Hong Kong The Hong Kong University of Science and Technology Shenzhen Research Institute Shenzhen518057 China The Hong Kong University of Science and Technology Department of Electronic and Computer Engineering The Department of Computer Science and Engineering Hong Kong Hong Kong

This study investigates barely-supervised medical image segmentation where only few labeled data, i.e., single-digit cases are available. We observe the key limitation of the existing state-of-the-art semi-supervised solution cross pseudo supervision is the unsatisfactory precision of foreground classes, leading to a degenerated result under barely-supervised learning. In this paper, we propose a novel Compete-to-Win method (ComWin) to enhance the pseudo label quality. In contrast to directly using one model's predictions as pseudo labels, our key idea is that high-quality pseudo labels should be generated by comparing multiple confidence maps produced by different networks to select the most confident one (a compete-to-win strategy). To further refine pseudo labels at near-boundary areas, an enhanced version of ComWin, namely, ComWin + , is proposed by integrating a boundary-aware enhancement module. Experiments show that our method can achieve the best performance on three public medical image datasets for cardiac structure segmentation, pancreas segmentation and colon tumor segmentation, respectively. The source code is now available at https://***/Huiimin5/comwin. © 1982-2012 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Surface Deposition of Ni(OH)_(2) and Lattice Distortion Induce the Electrochromic Performance Decay of NiO Films in Alkaline Electrolyte

引用

Energy & Environmental Materials 2024年第3期7卷 257-267页

作者： Kejun Xu Liuying Wang Chaoqun Ge Long Wang Bin Wang Zhuo Wang Chuanwei Zhang Gu Liu Xi'an Research Institute of High Technology Xi'an 710025China School of Material Science and Engineering Zhengzhou UniversityZhengzhou 450001China School of Mechanical Engineering Xi'an University of Science and TechnologyXi'an 710054China

NiO,an anodic electrochromic material,has applications in energy-saving windows,intelligent displays,and military ***,its electrochromic mechanism and reasons for its performance degradation in alkaline aqueous electrolytes are complex and poorly understood,making it challenging to improve NiO thin *** studied the phases and electrochemical characteristics of NiO films in different states(initial,colored,bleached and after 8000 cycles)and identified three main reasons for performance ***,Ni(OH)_(2)is generated during electrochromic cycling and deposited on the NiO film surface,gradually yielding a NiO@Ni(OH)_(2)core-shell structure,isolating the internal NiO film from the electrolyte,and preventing ion ***,the core-shell structure causes the mode of electrical conduction to change from first-to second-order conduction,reducing the efficiency of ion transfer to the surface Ni(OH)_(2)***,Ni(OH)_(2)and NiOOH,which have similar crystal structures but different b-axis lattice parameters,are formed during electrochromic cycling,and large volume changes in the unit cell reduce the structural stability of the thin ***,we clarified the mechanism of electrochromic performance degradation of NiO films in alkaline aqueous electrolytes and provide a route to activation of NiO films,which will promote the development of electrochromic technology.

关键词： alkaline electrolyte electrochromism NiO film performance attenuation mechanism

来源：评论

学校读者我要写书评

暂无评论

A charge system search based watermarking technique for gray-scale images using DWT–SVD

引用

Multimedia Tools and Applications 2024年第18期83卷 53743-53757页

作者： Gupta, Ritu Singh, Abhilasha Department of Information Technology Bhagwan Parshuram Institute of Technology GGSIPU Delhi India Department of Computer Science and Engineering SRM Institute of Science and Technology Delhi-NCR Campus Modinagar Ghaziabad India

This paper put forward an embedded scheme to execute image watermarking in light of the discrete wavelet transform (DWT), singular value decomposition (SVD) and Charge System Search (CSS) method. In the proposed scheme, Charge System Search technique is applied to advance the multiple scaling factors (MSFs). To approve the proposed scheme, six different images are used to inspect its performance and its outcomes are likewise contrasted with the past work announced in writing. From the results, it very well may be seen that proposed watermarking technique is a vigorous and proficient methodology in the field of grayscale image watermarking technology. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2023.

关键词： Discrete wavelet transforms

来源：评论

学校读者我要写书评

暂无评论

Short Term Traffic Flow Prediction Using Hybrid Deep Learning

引用

computers, Materials & Continua 2023年第4期75卷 1641-1656页

作者： Mohandu Anjaneyulu Mohan Kubendiran School of Computer Science and Engineering(SCOPE) Vellore Institute of Technology(VIT)Vellore632014India

Traffic flow prediction in urban areas is essential in the IntelligentTransportation System (ITS). Short Term Traffic Flow (STTF) predictionimpacts traffic flow series, where an estimation of the number of vehicleswill appear during the next instance of time per hour. Precise STTF iscritical in Intelligent Transportation System. Various extinct systems aim forshort-term traffic forecasts, ensuring a good precision outcome which was asignificant task over the past few years. The main objective of this paper is topropose a new model to predict STTF for every hour of a day. In this paper,we have proposed a novel hybrid algorithm utilizing Principal ComponentAnalysis (PCA), Stacked Auto-Encoder (SAE), Long Short Term Memory(LSTM), and K-Nearest Neighbors (KNN) named PALKNN. Firstly, PCAremoves unwanted information from the dataset and selects essential ***, SAE is used to reduce the dimension of input data using onehotencoding so the model can be trained with better speed. Thirdly, LSTMtakes the input from SAE, where the data is sorted in ascending orderbased on the important features and generates the derived value. Finally,KNN Regressor takes information from LSTM to predict traffic flow. Theforecasting performance of the PALKNN model is investigated with OpenRoad Traffic Statistics dataset, Great Britain, UK. This paper enhanced thetraffic flow prediction for every hour of a day with a minimal error *** extensive experimental analysis was performed on the benchmark *** evaluated results indicate the significant improvement of the proposedPALKNN model over the recent approaches such as KNN, SARIMA, LogisticRegression, RNN, and LSTM in terms of root mean square error (RMSE)of 2.07%, mean square error (MSE) of 4.1%, and mean absolute error (MAE)of 2.04%.

关键词： Short term traffic flow prediction principal component analysis stacked auto encoders long short term memory k nearest neighbors:intelligent transportation system

来源：评论

学校读者我要写书评

暂无评论

The IoT resource allocation and scheduling using Elephant Herding Optimization (EHO-RAS) in IoT environment

引用

International Journal of Information technology (Singapore) 2024年第5期16卷 3283-3293页

作者： Mageswari, Umaa Deepak, Gerard Santhanavijayan, A. Mala, C. Department of Computer Science and Engineering National Institute of Technology Tamilnadu Trichy 620015 India Department of Computer Science and Engineering Manipal Institute of Technology Karnataka Bengaluru 560064 India

IoT is one of the most significant technological breakthroughs and promises a higher level of connection and control in the future. The IoT network continues to expand rapidly, and the IoT ecosystem comprises millions of interconnected ad hoc devices across the network. Effective resource utilization guarantees the improvement of service quality. Everything is connected to the Internet through the distribution system known as the Internet of Things (IoT). Plenty of gateways and resources are in IoT infrastructure. Resource allocation (RA) is challenging due to network heterogeneity and the diversity of IoT devices;numerous practical approaches, strategies, and implementations are being presented and employed to resolve the RA problem (RAP). IoT resource allocation and scheduling (RAS) performance is essential in such a system since RAS allocates resources to open gateways and handles mapping resources and gateways. A gateway is needed to connect to hundreds of resources in the IoT environment. The proposed work is based on the RAS problem and aims to achieve optimal RA in the IoT by using the Elephant Herding Optimization (EHO) algorithm to lower the total Communication Cost between gateways and resources. The proposed EHO algorithm has been contrasted with others already in use, and the results show that the suggested algorithm performs as expected. The proposed solution is superior to others regarding TCC and Convergence rate than Particle Swarm Optimization (PSO), Genetic Algorithm (GA), Ant Colony Optimization (ACO), and Gray Wolf Optimization (GWO). © The Author(s), under exclusive licence to Bharati Vidyapeeth's institute of computer Applications and Management 2024.

关键词： Communication cost EHO Fitness IoT Gateways IoT Resources Resource Allocation ToT

来源：评论

学校读者我要写书评

暂无评论

Conv-attention ViT for classification of multi-label class imbalanced data of lung thoracic diseases

引用

Multimedia Tools and Applications 2024年 1-30页

作者： Oommen, Lintu Nikhila Nagajyothi, Chiluka Chebrolu, Srilatha Department of Computer Science and Engineering National Institute of Technology Andhra Pradesh Tadepalligudem India

Chest X-ray image classification is a key study topic, and in order to increase performance and accuracy, the efficiency of vision transformers for this task has been examined. However, imbalanced datasets pose a significant challenge during model training, leading to biased results. The main focus is the classification of multi-labeled lung thoracic diseases using imbalanced data. To tackle class imbalance, synthetic data was generated using GANs and combined with images from diverse data sources, which enhanced generalization. To classify multi-labeled chest X-rays, this work proposes a Conv-Attention Vision Transformer (CA-ViT) model that uses local and global attention mechanisms to improve the average ROC-AUC score and F1-score. Proposed model, CA-ViT’s performance was evaluated on an imbalanced dataset and compared with ResNet-50, VGG-19, and ViT 32/384 models. The studies demonstrated that CA-ViT surpassed the performance of the other models, scoring an average ROC-AUC of 0.76 as opposed to 0.72, 0.75, and 0.68 for ResNet-50, VGG-19, and ViT 32/384 models, respectively. All models displayed improved assessment metrics when the dataset was balanced via synthetic data creation and the combination of several data distributions. When compared to the other models, CA-ViT performed best, earning an average ROC-AUC of 0.81. In contrast, the average ROC-AUC values for the ResNet-50, VGG-19, and ViT 32/384 models were 0.7, 0.73, and 0.79 respectively. CA-ViT achieved micro-average F1-score of 0.70. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

Automatic Image Annotation Using Adaptive Convolutional Deep Learning Model

引用

Intelligent Automation & Soft Computing 2023年第4期36卷 481-497页

作者： R.Jayaraj S.Lokesh Department of Computer Science and Engineering Hindusthan College of Engineering and TechnologyCoimbatore641032TamilnaduIndia Department of Computer Science and Engineering Hindusthan Institute of TechnologyCoimbatore641032TamilnaduIndia

Every day,websites and personal archives create more and more *** size of these archives is *** comfort of use of these huge digital image gatherings donates to their ***,not all of these folders deliver relevant indexing *** the outcomes,it is dif-ﬁcult to discover data that the user can be absorbed ***,in order to determine the signiﬁcance of the data,it is important to identify the contents in an informative *** annotation can be one of the greatest problematic domains in multimedia research and computer ***,in this paper,Adap-tive Convolutional Deep Learning Model(ACDLM)is developed for automatic image ***,the databases are collected from the open-source system which consists of some labelled images(for training phase)and some unlabeled images{Corel 5 K,MSRC v2}.After that,the images are sent to the pre-processing step such as colour space quantization and texture color class *** pre-processed images are sent to the segmentation approach for efﬁcient labelling technique using J-image segmentation(JSEG).Theﬁnal step is an auto-matic annotation using ACDLM which is a combination of Convolutional Neural Network(CNN)and Honey Badger Algorithm(HBA).Based on the proposed classiﬁer,the unlabeled images are *** proposed methodology is imple-mented in MATLAB and performance is evaluated by performance metrics such as accuracy,precision,recall and F1_*** the assistance of the pro-posed methodology,the unlabeled images are labelled.

关键词： Deep learning model J-image segmentation honey badger algorithm convolutional neural network image annotation

来源：评论

学校读者我要写书评

暂无评论

A transfer learning approach for continuous speech recognition system in Indian language sadri

引用

International Journal of Information technology (Singapore) 2025年 1-10页

作者： Paul, Shubhojeet Bhattacharjee, Vandana Saha, Sujan Kumar Department of Computer Science and Engineering Birla Institute of Technology Mesra Jharkhand Ranchi 835215 India Department of Computer Science and Engineering National Institute of Technology West Bengal Durgapur 713209 India

Sadri is the most widely used language of the Chotanagpur Plateau region of India. This is primarily a spoken language and developing an automatic speech recognition (ASR) system in Sadri is extremely important. When we searched the literature, we found no open ASR system or relevant resources in Sadri. So, we worked on developing an ASR system in Sadri. We created a Sadri speech corpus of around 20 h, employing 53 native speakers. Then, we implemented the baseline model using Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN). As the size of the training data was not sufficient, we investigated a transfer learning technique to improve performance. We found that ASR data are openly available in other related languages such as Hindi and Bengali. We used Bengali and Hindi data of 100 h each in a transfer learning framework along with the Sadri data. In our experiments, we found that the transfer learning-based model outperforms the baseline model. © Bharati Vidyapeeth's institute of computer Applications and Management 2025.

关键词： ASR Deep learning Indian language Sadri Speech recognition

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：