检索结果-内蒙古大学图书馆

48th IEEE international conference on Acoustics, Speech and signal processing, ICASSP 2023

作者： Fan, Wenkang Zhang, Kaiyun Shi, Hong Chen, Jianhua Chen, Yinran Luo, Xiongbiao Xiamen University National Institute for Data Science in Health and Medicine Xiamen361102 China Fujian Medical University Cancer Hospital Fujian Cancer Hospital Fuzhou350014 China Xiamen University Department of Computer Science Xiamen361005 China

ISBN: (纸本)9781728163277

Surface reconstruction is an essential way to expand surgical field of view during endoscopic surgery, but it certainly requires dense depth estimation of endoscopic video sequences. Unfortunately, such a dense depth recovery suffers from illumination variation, weak texture, and occlusion. To address these problems, this work proposes a new triple-supervision self-learning strategy that uses unannotated endoscopic video data to predict monocular endoscopic dense depth information. This strategy first employs an effective conventional method to estimate camera poses and sparse depth maps to establishing a sparse data self-supervision. Furthermore, our strategy still combines two consistency measures to supervise dense depth and photometric information. We evaluated our method on collected colonoscopic videos, with the experimental results showing that our triple-supervision learning framework works more effective and accurate than some current self-supervised and unsupervised learning methods. © 2023 IEEE.

关键词： Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

A remote sensing work-flow recommendation algorithm based on logical structure

A remote sensing work-flow recommendation algorithm based on...

引用

2022 international conference on signal processing, Computer Networks, and Communications, SPCNC 2022

作者： Zuo, Xianyu Wu, Yaohua Liu, Yang Lin, Yinghao Zhang, Lei Henan Key Laboratory of Big Data Analysis and Processing School of Computer and Information Engineering Henan University Kaifeng475001 China

ISBN: (纸本)9781510664616

With the development of remote sensing products towards the direction of civilianization and popularization, work-flow customization plays an important role in the production of remote sensing products. The traditional customized work-flow model has large time cost, complex operation and high requirements for professional knowledge. The work-flow recommendation system can improve the construction efficiency of remote sensing work-flow to some extent and assist users to design high-quality remote sensing work-flow models. However, most of the existing remote sensing work-flow modeling methods ignore the logical structure characteristics of the work-flow, leading to large errors in the calculation results of similarity. Difficult to make work-flow recommendations effectively. Therefore, this paper proposes a customized recommendation algorithm for remote sensing work-flow based on logical structure. By focusing on logical structure, the reliability of similarity calculation between work-flow is improved, so as to find similar work-flow to help users recommend the next modeling node. Firstly, the work-flow model needs to be preprocessed: the user converts the constructed work-flow model into a process structure tree by using Petri net workflow, and uses the path table generation algorithm based on logical structure to convert the model information into data information and store it in the database for subsequent data processing;Then the flow data in the flow tree set was converted into a path table according to certain rules, and then the longest common subsequence similarity of each data in the path table was calculated to obtain the similarity calculation results based on the logical structure characteristics, the most similar work-flow in the work-flow library is found and the recommendation is made for the user. The method proposed in this paper is evaluated experimentally on the real data set, in terms of recall, precision and F1-score, which shows that the method

关键词： Remote sensing

来源：评论

学校读者我要写书评

暂无评论

An approach towards AGV battery cell voltage prediction using DL and data pre-processing 18

An approach towards AGV battery cell voltage prediction usin...

引用

18th IEEE international conference on Computer Science and Information Technologies, CSIT 2023

作者： Medykovskyy, Mykola Cupek, Rafal Pavliuk, Olena Mishchuk, Myroslav Lviv Polytechnic National University ACS Department Lviv Ukraine Silesian University of Technology Department of Distributed Systems and Informatic Devices Gliwice Poland

ISBN: (纸本)9798350360462

Modern methods for solving the AGV battery cell voltage prediction problem include a symbiosis of probabilistic and machine learning methods. In this study, we propose to use two RNN-based approaches with preliminary data analysis and processing. In the first approach, we make predictions based on smoothed data. Three filters were chosen for smoothing: exponential smoothing;moving average;weighted moving average. The moving average method showed the best results, but according to the Dickey-Fuller test, the predicted signal was still not stationary. The second approach was based on a centered stationary signal obtained by subtracting the value of the moving average from the original *** forecast accuracy was evaluated using the following errors: MAPE, RMSE, MSPE, RMSPE, MFE, SEBIAS, ME. Adequacy of the proposed neural network model based on the coefficient of determination R2. As can be seen from Table 2, a significant improvement in prediction accuracy is observed for the centered signal. The maximum error was reduced by 5.6 times R2 increasing by 0.0596. At the same time, both models are adequate. © 2023 IEEE.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Generating Fake Cyber Threat Intelligence Using the GPT-Neo Model 8

Generating Fake Cyber Threat Intelligence Using the GPT-Neo ...

引用

8th international conference on Intelligent Computing and signal processing, ICSP 2023

作者： Song, Zheyu Tian, Yonghong Zhang, Junjin Hao, Yufeng Inner Mongolia University of Technology College of Data Science and Application Hohhot China

ISBN: (纸本)9798350302455

At present, some network security defense systems expand their knowledge base by extracting network threat Intelligence (CTI) to understand the common attack techniques and processes of malicious attack groups. However, there are potential risks to this approach. Hackers can spread fake CTI through the Open-Source Intelligence (OSINT) platform to fool defense systems into learning false information. In this article, we focus on how to generate fake CTI text using the GPT-Neo model and show that the generated text is highly confident. By fine-tuning a common language model like GPT-Neo, we can generate text similar to real CTI. At the end of the paper, we propose two disinformation detection methods that can help us to eliminate unreliable content. © 2023 IEEE.

关键词： Network security

来源：评论

学校读者我要写书评

暂无评论

A novel coriolis mass flowmeter signal processing method based on adaptive notch filter and Hilbert transform

引用

ENERGY REPORTS 2022年 8卷 265-272页

作者： Shen Ting'ao Chen Liwei Huang Siyu Zhang Peili Army Logist Acad Chongqing 401311 Peoples R China Sichuan Taili Engn Project Management Co Ltd Chengdu 610023 Peoples R China

Coriolis mass flowmeter has been widely used in the energy field, and the key of mass flowrate measurement is to calculate the frequency and phase difference of the signal accurately. Aiming at the poor measurement precision problem in Coriolis mass flowmeter, a novel Coriolis mass flowmeter signal processing method is presented based on adaptive notch filter and Hilbert transform. Firstly, the long time continuous frequency tracking and accuracy problem of adaptive notch filter is effectively solved, by introducing the idea of negative feedback control. Then, the frequency estimation results are used to judgment whether the whole data will be extended or not, which makes the length of data sequence close to or equal to an integer multiple of the period. And then, Hilbert transform is conducted to the extended signals. Finally, the phase difference can be obtained by utilizing the sinusoidal functions. Simulation results show that the proposed method has high frequency and phase difference estimation accuracy, and it can be used for real time flowrate measurement of flowmeter. Experimental tests validate that the precision of this method is less than 3h, which is more than 2 times of that of the traditional method. O 2022 Published by Elsevier Ltd. This is an open access article under theCCBY-NC-ND license (http://***/licenses/by-nc-nd/4.0/).

关键词： Hilbert transform Adaptive notch filter Parameter estimation Coriolis mass flowmeter Flowrate measurement

来源：评论

学校读者我要写书评

暂无评论

Feature Extraction for Machine Learning-based Sleep Stage Classification Using PPG-Derived Parameters and Skin Temperature

Feature Extraction for Machine Learning-based Sleep Stage Cl...

引用

2024 Asia Pacific signal and Information processing Association Annual Summit and conference, APSIPA ASC 2024

作者： Buaruk, Suphachok Thanaviratananich, Sikawat Treesuthacheep, Peerasit Deepaisarn, Somrudee Sirindhorn International Institute of Technology Thammasat University Pathum Thani Thailand Division of Neurology Department of Medicine Bumrungrad International Hospital Bangkok Thailand Sleep Disorders Center Cleveland Clinic Neurological Institute ClevelandOH United States Chulalongkorn Comprehensive Epilepsy Center of Excellence King Chulalongkorn Memorial Hospital Bangkok Thailand Research Unit in Sustainable Electrochemical Intelligent Thammasat University Pathum Thani Thailand

ISBN: (纸本)9798350367331

This study explores feature extraction and machine learning methods for automated sleep stage classification using physiological signals, focusing on PPG-derived parameters and skin temperature. We experimented with feature extraction techniques that required different engineering steps, i.e., with or without applying the first derivative and deviation from the mean of raw signals. The statistical values, including mean, standard deviation, maximum, minimum, and median of each 30-second segment of signals, are then extracted and used as input features for our selected classification model, the Extra Trees. The classification performance is evaluated under both subject-dependent and subject-independent scenarios using data from 93 subjects. The data include labeled stages of wake, non-rapid eye movement (NREM) sleep stages N1, N2, and N3, as well as rapid eye movement (REM) sleep. Optimal features were derived from skin temperature measurement along with the normalized timestamp, which is associated with sleep duration. Results show that the subject-dependent experiment yields the highest performance with an accuracy of 0.87 for three-stage classification (wake, NREM:N1/N2/N3, REM) and 0.77 for five-stage classification (wake, N1, N2, N3, REM). This work highlights the potential of artificial intelligence to enable the automatic labeling of sleep stages using non-constrained sensors. The important features related to sleep staging can be investigated, which could contribute to future advances in clinical diagnostics and healthcare applications. © 2019 Institute of Physics Publishing. All rights reserved.

关键词： Sleep research

来源：评论

学校读者我要写书评

暂无评论

An Efficient Client Clustering Algorithm for Clustered Federated Learning

An Efficient Client Clustering Algorithm for Clustered Feder...

引用

IEEE/CIC international conference on Communications in China (ICCC)

作者： Su, Junshen Wang, Xijun Chen, Xiang Sun Yat Sen Univ Sch Elect & Informat Technol Guangzhou Peoples R China Tsinghua Univ Shenzhen RITS Res Inst Shenzhen Peoples R China Guangxi Key Lab Wireless Wideband Commun & Signal Guilin Peoples R China

ISBN: (数字)9781665484800

ISBN: (纸本)9781665484800

Federated Learning (FL) is a promising distributed machine learning framework that trains a global model with privacy-preserving. data heterogeneity is one of the biggest challenges in FL. It has been shown that concept shift leads to suboptimal performance with a single global model. To address this issue, we employ Clustered Federated Learning (CFL), which is to divide clients into different groups according to their local data distributions. In this paper, we propose a cosine similarity-based clustering algorithm without modifying the conventional FL communication protocol. In particular, our algorithm does not require the number of groups to be known in advance. Furthermore, we present a client scheduling scheme that can accelerate the convergence of the clustering algorithm. Experiments show that our clustering algorithm can effectively cluster clients and significantly improve the test accuracy compared to FL without clustering. We also demonstrate that our client scheduling scheme allows the clustering algorithm to converge in fewer communication rounds.

关键词： Federated learning non-IID clustering concept shift

来源：评论

学校读者我要写书评

暂无评论

FPGA-Based Implementation of Cuffless Blood Pressure Measurement Using Photoplethysmogram signal 3rd

FPGA-Based Implementation of Cuffless Blood Pressure Measur...

引用

3rd international conference on signal and data processing, icsdp 2023

作者： Yudha Raja Singam, D. Mukhopadhyay, Sumitra Gupta, Rajarshi Institute of Radio Physics and Electronics University of Calcutta Kolkata India Electrical Engineering Applied Physics University of Calcutta Kolkata India

ISBN: (纸本)9789819795772

In recent years, there has been research on non-invasive blood pressure (NIBP) measurement using physiological signals. However, the main challenges are the selection of physiological signals, low complexity algorithm development and its hardware realization. This paper focuses on FPGA-based realization of a NIBP measurement model using Photoplethysmogram (PPG) signal. The PPG signal is pre-processed, beats are segmented, autoencoder features are extracted, and among 12 different regression models the most competitive model is selected to predict the systolic (SBP) and diastolic blood pressure (DBP) from PPG beat. The model was evaluated with PhysioNet MIMIC-III waveform database and achieved mean absolute error (MAE) and standard deviation (SD) of 3.69 and 5.06 mmHg respectively for SBP, and 1.96 and 3.68 mmHg for DBP respectively. The model has been implemented into Zynq-7000 SoC ZC702 standalone hardware with the lower detection latency (∼6.24 ms for 1 PPG beat) reported so far in the literature and low memory requirement (∼78 KB). © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Blood pressure

来源：评论

学校读者我要写书评

暂无评论

A New Personalized Efficacy Atlas for Pallidal Deep Brain Stimulation 48

A New Personalized Efficacy Atlas for Pallidal Deep Brain St...

引用

48th IEEE international conference on Acoustics, Speech and signal processing, ICASSP 2023

作者： Luo, Xiongbiao Xiamen University Department of Computer Science Xiamen36105 China Xiamen University National Institute for Data Science in Health and Medicine Xiamen361102 China

ISBN: (纸本)9781728163277

Neurostimulation is to implant implants electrodes into deep brain structures to treat drug-resistant motor disorder in Parkinson's disease. Unfortunately, it remains challenging to find the optimal electrode implanted and activated location at deep anatomical regions and achieve the optimal surgical function or performance on patients. This paper proposes to create a novel personalized efficacy atlas that warps functional scales and estimates activation volume by modeling electric field to characterize the link between electrode location and neurosurgical performance. We used a population of 32 globus pallidus stimulated patient data to construct such an atlas. The experimental results demonstrate that modeling electric field centered at actually electrode implanted positions outperforms Gaussian kernel modeling to predict activation volume for functional atlas creation. In particular, our new atlas provides an automatic and accurate electrode implantation method with a guidance accuracy 1.26 mm, which is better than other approaches. Additionally, two phases of the functional scales obtained from 3 and 6 months after neurostimulation were compared to create the new atlas. The 6-month phase gives a better efficacy map. © 2023 IEEE.

关键词： Deep brain stimulation electric field modeling functional atlas volume of tissue activated

来源：评论

学校读者我要写书评

暂无评论

PRIOR-BERT AND MULTI-TASK LEARNING FOR TARGET-ASPECT-SENTIMENT JOINT DETECTION 47

PRIOR-BERT AND MULTI-TASK LEARNING FOR TARGET-ASPECT-SENTIME...

引用

47th IEEE international conference on Acoustics, Speech and signal processing (ICASSP)

作者： Ke, Cai Xiong, Qingyu Wu, Chao Liao, Zikai Yi, Hualing Chongqing Univ Sch Big Data & Software Engn Chongqing Peoples R China Chongqing Univ Key Lab Dependable Serv Comp Cyber Phys Soc Minist Educ Chongqing Peoples R China Xidian Univ Xian Peoples R China

ISBN: (纸本)9781665405409

Aspect-Based Sentiment Analysis (ABSA) is a fine-grained sentiment analysis task and has become a significant task with real-world scenario value. The challenge of this task is how to generate an effective text representation and construct an end-to-end model that can simultaneously detect (target, aspect, sentiment) triples from a sentence. Besides, the existing models do not take the heavily unbalanced distribution of labels into account and also do not give enough consideration to long-distance dependence of targets and aspect-sentiment pairs. To overcome these challenges, we propose a novel end-to-end model named Prior-BERT and Multi-Task Learning (PBERT-MTL), which can detect all triples more efficiently. We evaluate our model on SemEval-2015 and SemEval-2016 datasets. Extensive results show the validity of our work in this paper. In addition, our model also achieves higher performance on a series of subtasks of target-aspect-sentiment detection. Code is available at https://***/CQUPTCaiKe/PBERT-MTL.

关键词： Aspect-Based Sentiment Analysis Target-Aspect-Sentiment Multi-Task Learning Joint Detection

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：