检索结果-内蒙古大学图书馆

2022 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2022

作者： Bai, Tian Zhang, Yingfang Wang, Yuzhao Qin, Yanguo Zhang, Fa Jilin University College of Computer Science and Technology Changchun China Jilin University Key Laboratory of Symbolic Computation and Knowledge Engineering Ministry of Education Changchun China The Second Hospital of Jilin University The Department of Orthopaedics Changchun China Beijing Institute of Technology School of Medical Technology Beijing China

ISBN: (纸本)9781665468190

Deep learning often requires large amounts of data from different institutions. Federated learning, as a distributed training framework, enables multiple participants to collaboratively train models without collecting data together and hence protecting data privacy, but the datasets from different institutions usually bring the problem of domain shift, which affects the performance of the model. When addressing domain shift, previous works often use a single global model to share parameters. Therefore, we propose a novel method to train multiple public models with different structures under the federated framework to improve the reliability and robustness of the public models. And each participant keeps its own domain-tuned private model, the private model does not share parameters with other participants. We use Mixture of Experts (MoE) domain adaptation to dynamically combine different public models and private model, which utilizes the similarity between different datasets to update the parameters of the public models. We apply the proposed method to the multi-site Magnetic resonance imaging (MRI) end-to-end classification, and the experiments demonstrate its effectiveness. © 2022 IEEE.

关键词： Magnetic resonance imaging

来源：评论

学校读者我要写书评

暂无评论

An efficient and real-time steel surface defect detection method based on single-stage detection algorithm

引用

Multimedia Tools and Applications 2024年第42期83卷 90595-90617页

作者： Zhang, Hongkai Miao, Qiqi Li, Suqiang Wang, Chunping Chan, Sixian Hu, Jie Bai, Cong School of Electronic and Information Engineering Anhui Jianzhu University Ziyun Road Anhui Province Hefei230601 China Key Laboratory for Comprehensive Energy Saving of Cold Regions Architecture of Ministry of Education Jilin Jianzhu University Xincheng Street Jilin Province Changchun130119 China College of Computer Science and Technology Zhejiang University of Technology Liuhe Street Zhejiang Province Hangzhou310023 China Key Laboratory of Intelligent Informatics for Safety & Emergency of Zhejiang Province Wenzhou University Meiquan Street Zhejiang Province Wenzhou325035 China

Within the realm of computer vision, the detection of defects holds significant importance for industrial production and quality assurance. However, current object detection methods encounter notable challenges when tasked with defect detection, including high inter-class similarity and substantial intra-class variability. These challenges restrict the performance of conventional approaches, leading to imprecise defect detection and localization. In response to these limitations, this study proposes a novel real-time, end-to-end defect detection approach based on a single-stage detection algorithm. Our method introduces several innovations to address these challenges and significantly enhance defect detection capabilities. Firstly, we develop a pioneering cross-layer feature aggregation method, enabling the network to effectively capture intricate defect patterns and improve its adaptability to defect detection tasks. Additionally, we introduce an attention-enhanced decoupled head, which prioritizes critical features while optimizing computational efficiency. Through comprehensive testing on the NEU-DET defect detection dataset, our approach demonstrates remarkable performance gains over conventional object detection methods. Notably, our method achieves an impressive detection speed of 68 FPS while surpassing existing methods in terms of detection accuracy. These results highlight the practical significance of our approach, meeting the stringent requirements of industrial production for both accuracy and speed. Furthermore, the efficacy and robustness of our method are further validated using the GC10-DET dataset, underscoring its potential for real-world applications across diverse industrial settings. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Prognostics of rolling element bearing based on combined SSA-VMD and CNN-BiGRU models

Prognostics of rolling element bearing based on combined SSA...

引用

Fault Detection, Supervision and Safety for Technical Processes (SAFEPROCESS), CAA Symposium on

作者： Bangcheng Zhang Jingyuan Song Bo Li Zhi Gao School of Mechatronics Changchun Institute of Engineering and Changchun University of Technology Changchun China Jilin Emergency Management Institute Changchun Institute of Engineering Changchun China School of Computer Science and Computing Changchun Institute of Engineering Changchun China School of Applied Technology Changchun University of Technology Changchun China

In industrial production, rolling element bearings are key rotating components in a wide range of applications. By predicting rolling element bearings, the failure of bearings as well as problems can be detected in time and preventive measures can be taken to improve production efficiency before the failure occurs. For the prediction problem of rolling element bearing, a hybrid model of rolling bearing prediction based on variable mode decomposition optimized by sparrow search algorithm and bidirectional recurrent neural network of convolutional neural network based on attention mechanism is proposed. Firstly, the initial signal is preprocessed with data using variational modal decomposition. Secondly, the signal with threshold value more than 0.5 is selected as the input of convolutional neural network to perform feature extraction. Then, the bidirectional recurrent neural network structure is used to perform the degradation prediction of rolling bearings. Finally, the proposed method is validated using the XJTU-SY dataset and compared with gated recurrent unit, bidirectional recurrent neural network, bidirectional recurrent neural network based on attention and attention-based convolutional neural network, bidirectional recurrent neural network combination model. The experimental results show that the used model has about 30% higher prediction accuracy than the above four models and has some application val.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Iterative learning algorithm of upper limb rehabilitation robot based on BP neural network

Iterative learning algorithm of upper limb rehabilitation ro...

引用

International Conference on Intelligent Control and Information Processing (ICICIP)

作者： Zaixiang Pang Xiaomeng Deng Linan Gong Nan Wang Changchun University of Technology School of Engineering training center Changchun China Changchun University of Technology School of Computer science and engineering Changchun China Changchun Vocational Institute of Technology School of Mechanical and Electrical Engineering Changchun China Jilin Weimiao Medical Technology Co. LTD Changchun China

At present, in the trajectory tracking control of upper limb rehabilitation robot, there are nonlinear and uncertain problems of patients' spasticity disturbance. Therefore, an iterative learning control algorithm based on Back Propagation neural network is proposed in this paper. Based on the repeatability of the rehabilitation robot system, a multi-layer feedforward neural network trained by error reverse propagation algorithm is combined with iterative learning. The algorithm can effectively compensate the effect of spasmodic interference on the system performance and improve the convergence rate of trajectory tracking errors. The structure of the iterative learning control system based on neural network and the iterative learning control algorithm based on Back Propagation neural network are described. The process of network selection and robot trajectory tracking iterative learning control is given. Through simulation experiments, the feasibility and effectiveness of the algorithm are verified, which solves the above problems and meets the control requirements of rehabilitation training.

关键词：

来源：评论

学校读者我要写书评

暂无评论

American Sign Language Recognition Based on Machine Learning and Neural Network 2

American Sign Language Recognition Based on Machine Learning...

引用

2nd International Conference on Machine Learning and Intelligent Systems engineering, MLISE 2022

作者： Li, Lanxi Liu, Da Shen, Chenlin Sun, Jing The Ohio State University Computer Science And Engineering Columbus43201 United States Sichuan University Department Of Computer Science Chengdu610207 China Jilin University Faculty Of Software Engineering Changchun130012 China Institute Of Information And Software Engineering University Of Electronic Science And Technology Of China Chengdu610066 China

ISBN: (数字)9781665492461

ISBN: (纸本)9781665492461

Numerous disabilities such as deaf and mute are suffered from not being capable of communicating with normal people, it is necessary to find a way to solve this problem. A feasible method is Sign Language Recognition (SLR) which is a sort of pattern recognition technique. In this paper, machine learning and deep learning methods are applied to recognize and classify American Sign Language (ASL), and only 24 English letters are classified because letter J and Z require fingers to move. First, Principal Component Analysis (PCA) and manifold algorithms are used to do dimension reduction to accelerate the training of machine learning and visualize it. Second, various machine learning methods such as Random Forest Classification (RFC), K-Nearest Neighbor (KNN), Gaussian Naïve Bayes (GNB), Support Vector Machine (SVM), and Stochastic Gradient Descent (SGD) are applied to classify the pattern. Since the SVM algorithm has several hyperparameters, this study uses the Grid Search method to find the best combination of hyperparameter which lead to predicting more accurately. It is found that different dimensionality reduction algorithms have unequal effects on the accuracy of each prediction model, and it can be concluded that the manifold algorithm is the best dimension reduction algorithm only for KNN but not for other prediction models, and PCA is much more feasible than KNN applied in such machine learning algorithms except KNN. Two deep learning algorithms such as Convolutional Neural Networks (CNN) and Deep Neural Networks (DNN) are also used in classification and their accuracy is highest among such algorithms mentioned above. © 2022 IEEE.

关键词： Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

A Segmentation Framework for Acoustic Sidescan Sonar Images Using Improved Smallest of Constant False Alarm Rate and MAP-MRF 16

A Segmentation Framework for Acoustic Sidescan Sonar Images ...

引用

16th International Conference on Underwater Networks and Systems, WUWNet 2022

作者： Tang, Yiteng Liu, Jun Song, Shanshan Guan, Wenxue Cui, Jun-Hong College of Computer Science and Technology Jilin University Jilin Province Changchun China School of Electronic and Information Engineering Beihang University Haidian Qu Beijing Shi China Smart Ocean Science and Technology Research Institute Jilin University Jilin Province Changchun China Shenzhen Institute for Advanced Study University of Electronic Science and Technology of China Guangdong Province Shenzhen China

ISBN: (纸本)9781450399524

Segmentation of sidescan sonar image is a significant issue in underwater object detection and recognition. However, most prior methods only consider segmentation accuracy, ignoring false alarm rate, which plays a vital role in object detection and recognition. In this paper, a robust and accurate segmentation framework for sidescan sonar image is proposed, which balances a preferred tradeoff between accuracy and false alarm rate. The proposed method integrates an improved Smallest Of Constant False Alarm Rate (SO-CFAR) algorithm and a Maximum A Posteriori probability and Markov Random Field model (MAP-MRF). The part of innovations segments acoustical highlight region accurately while preserving edge features, which can make segmentation results obtain preferred false alarm rate. After that, MAP-MRF is employed for overcoming drawbacks associated with higher threshold value in continuous acoustical highlight areas. Besides, to better deal with intensity inhomogeneity, Density-Based Spatial Clustering of Applications with Noise (DBSCAN) is incorporated into this method, which can locate Region Of Interest (ROI) in sonar images as well as improve segmentation effect. Experimental and comparative results on actual side-scan sonar images demonstrate that our method provides superior denoising, precision, and robustness performance. © 2022 ACM.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Improving Local Search for Pseudo Boolean Optimization by Fragile Scoring Function and Deep Optimization 29

Improving Local Search for Pseudo Boolean Optimization by Fr...

引用

29th International Conference on Principles and Practice of Constraint Programming, CP 2023

作者： Zhou, Wenbo Zhao, Yujiao Wang, Yiyuan Cai, Shaowei Wang, Shimao Wang, Xinyu Yin, Minghao School of Information Science and Technology Northeast Normal University Changchun China Key Laboratory of Applied Statistics of MOE Northeast Normal University Changchun China Key Laboratory of Symbolic Computation and Knowledge Engineering of MOE Jilin University Changchun China State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China

ISBN: (纸本)9783959773003

Pseudo-Boolean optimization (PBO) is usually used to model combinatorial optimization problems, especially for some real-world applications. Despite its significant importance in both theory and applications, there are few works on using local search to solve PBO. This paper develops a novel local search framework for PBO, which has three main ideas. First, we design a two-level selection strategy to evaluate all candidate variables. Second, we propose a novel deep optimization strategy to disturb some search spaces. Third, a sampling flipping method is applied to help the algorithm jump out of local optimum. Experimental results show that the proposed algorithms outperform three state-of-the-art PBO algorithms on most instances. © Wenbo Zhou, Yujiao Zhao, Yiyuan Wang, Shaowei Cai, Shimao Wang, Xinyu Wang, and Minghao Yin.

关键词： Combinatorial optimization

来源：评论

学校读者我要写书评

暂无评论

Partially training, isolating and unlearning, mitigating backdoor attack

Partially training, isolating and unlearning, mitigating bac...

引用

IEEE International Conference on Big Data

作者： Yong Li Han Gao Hao Chen Yiming Wang Chun Yu School of Computer Science and Engineering Changchun University of Technology AI Research Institute Changchun University of Technology School of Computer Science and Technology Jilin University Changchun China Non-Commissioned Officer School of Army Academy of Armored Forces Changchun China

ISBN: (数字)9798350362480

ISBN: (纸本)9798350362497

Neural networks are widely known to be vulnerable to backdoor attacks, witch are full of threats, a method that poisons a portion of the training data to make the target model perform well on normal data sets, while outputting attacker-specified or random categories on the poisoned samples. Nowadays, poisoned samples are becoming more and more similar to corresponding normal samples, even the human eye cannot easily distinguish them and the accuracy of models carrying backdoors on normal samples is no different from that of clean models witch also leads to the inability to determine whether a model carries a backdoor based on its accuracy. In this article, by observing the characteristics of backdoor attacks, We provide a new model training method, Partial Training(PT), that freezes part of the model to train a model that can isolate suspicious samples. Then, on this basis, a clean model is fine-tune to resist backdoor attacks. We used three datasets and four backdoor attack methods to conduct our experiments. The experiments showed that PT performed very well in defending against backdoor attacks mentioned above.

关键词： Training Accuracy Toxicology Neural networks Training data Resists Big Data Data models

来源：评论

学校读者我要写书评

暂无评论

Slow-light-enhanced on-chip 1D and 2D photonic crystal waveguide gas sensing in near-IR with an ultrahigh interaction factor

引用

Photonics Research 2023年第10期11卷 1647-1656页

作者： ZIHANG PENG YIJUN HUANG KAIYUAN ZHENG CHUANTAO ZHENG MINGQUAN PI HUAN ZHAO JIALIN JI YUTING MIN LEI LIANG FANG SONG YU ZHANG YIDING WANG FRANK K.TITTEL State Key Laboratory of Integrated Optoelectronics College of Electronic Science and EngineeringJilin UniversityChangchun 130012China Department of Electrical Engineering and Photonics Research Institute The Hong Kong Polytechnic UniversityHong Kong 518060China State Key Laboratory of Luminescence and Applications Changchun Institute of OpticsFine Mechanics and PhysicsChinese Academy of SciencesChangchun 130033China Department of Electrical and Computer Engineering Rice UniversityHoustonTexas 77005USA

Nanophotonic waveguides hold great promise to achieve chip-scale gas sensors. However, their performance is limited by a short light path and small light–analyte overlap. To address this challenge, silicon-based, slow-lightenhanced gas-sensing techniques offer a promising approach. In this study, we experimentally investigated the slow light characteristics and gas-sensing performance of 1D and 2D photonic crystal waveguides(PCWs) in the near-IR(NIR) region. The proposed 2D PCW exhibited a high group index of up to 114, albeit with a high propagation loss. The limit of detection(LoD) for acetylene(C_(2)H_(2)) was 277 parts per million(ppm) for a1 mm waveguide length and an averaging time of 0.4 s. The 1D PCW shows greater application potential compared to the 2D PCW waveguide, with an interaction factor reaching up to 288%, a comparably low propagation loss of 10 dB/cm, and an LoD of 706 ppm at 0.4 s. The measured group indices of the 2D and 1D waveguides are104 and 16, respectively, which agree well with the simulation results.

关键词： waveguide light interaction

来源：评论

学校读者我要写书评

暂无评论

Effects of Electron Donating Ability of Substituents and Molecular Conjugation on the Electronic Structures of Organic Radicals

引用

Chemical Research in Chinese Universities 2023年第2期39卷 202-207页

作者： LI Pengyuan FENG Lu LI Guangyue BAI Fuquan International Joint Research Laboratory of Nano-Micro Architecture Chemistry Institute of Theoretical Chemistry and College of ChemistryJilin UniversityChangchun130021P.R.China College of Chemical Engineering North China University of Science and TechnologyTangshan063210P.R.China International Innovation Center for Forest Chemicals and Materials Nanjing Forestry UniversityNanjing210037P.R.China

The geometries and electronic structures of a series of electron donor-acceptor radical molecules have been studied *** computational results show that the introduction of substituents with strong electron donating ability into tri-(2,4,6-trichlorophenyl)methyl(TTM)radicals enables the radical molecules to form the non-Aufbau electronic *** difficulty of forming the non-Aufbau electronic structure decreases with the enhancement of the electron donating ability of the substituent,but the expansion of the molecular conjugated system is not conducive to the *** hybridization of different fragments in molecular orbitals results in the disproportionation of orbital energy level and forms a staggered energy level *** electronic structure of radical molecules can be adjusted by substituents and molecular skeleton profoundly,which is a very effective means for molecular design.

关键词： Radical emitter Non-Aufbau electronic structure Molecular orbital characteristics Density functional theory and time-dependent density functional theory(DFT and TDDFT)

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：