检索结果-内蒙古大学图书馆

International Congress on image and Signal processing, BioMedical Engineering and Informatics

作者： Ting Feng Zuoyong Li Chuansheng Wang Xinwei Chen Jiawei Wu Fujian Provincial Key Laboratory of Information Processing and Intelligent Control Minjiang University Fuzhou China School of Computer Science and Technology Harbin University of Science and Technology Harbin China College of Mathematics and Computer Science Fuzhou University Fuzhou China

Hazy weather brings a lot of inconvenience to peoples lives, such as transportation. Therefore, image dehazing is still an important focus point. To achieve image dehazing, we proposed a Deep Dilated Residual Haze Network (DDRHNet) based on self-encoder and self-decoder. The proposed mode is an end-to-end supervised method, which directly estimates image dehazing result instead of estimating the atmospheric light and the transmission in the unsupervised and classical atmospheric scattering model. The DDRHNet method obtains image dehazed result by finding the difference between the input hazy image and the output of haze-free image. Experimental results on an open image dehazing dataset called SOTS demonstrate the superiority of the proposed DDRHNet network.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Lightweight Chinese Summarization Generation Method Based on Knowledge Distillation

A Lightweight Chinese Summarization Generation Method Based ...

引用

Information Science, Parallel and Distributed Systems (ISPDS), International Conference on

作者： Shiyue Liu Yannan Xiao Xueyong Wang Wenming Huang Zhenrong Deng Guangxi Key Laboratory of Image and Graphic Intelligent Processing School of Computer and Information Security Guilin University of Electronic Technology Guilin China School of Computer and Information Security Guilin University of Electronic Technology Guilin China Guilin Xintong Technology Co. Guilin China

In recent years, large-scale pretrained natural language processing models such as BERT, GPT3 have achieved good results in processing tasks. However, in daily applications, these large-scale language models usually exist large model size or long running time, and the model portability and application are not very convenient. To settle this problem, we come up with a lightweight summarization generation method based on knowledge distillation (LW-BERT-KD, abbreviated as LBK model) to prove that complex neural network knowledge (taking BERT model as an example) can be extracted to lightweight language processing models (taking BiLSTM model as an example), thus achieving the purpose of complex knowledge extraction and model compression. In this paper, the BERT model is regarded as the teacher model, and the BiLSTM is regarded as the student model. The knowledge distillation technology is the transfer of knowledge from teacher models to student models, so that the generated model is effective, lightweight, and easy to port. On the Chinese abstract generation dataset, the training effect of this model is significantly better than the basic Transformer baseline. The experimental results show that if the training parameters are increased by 100 times, the model achieves comparable results with complex preprocessing language models.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Early Forest Fire Segmentation Based on Deep Learning

Early Forest Fire Segmentation Based on Deep Learning

引用

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

作者： Mengna Li Youmin Zhang Jing Xin Lingxia Mu Ziquan Yu Han Liu Guo Xie Shangbin Jiao Yingmin Yi Shaanxi Key Laboratory of Complex System Control and Intelligent Information Processing Xi’an University of Technology Xi’an Shaanxi China Concordia University Montreal Quebec Canada College of Automation Engineering Nanjing University of Aeronautics and Astronautics Nanjing 211106 China

ISBN: (纸本)9781665401166

Fire segmentation is very important for fire rescue. It can make firefighters get the information on fire area, spread direction and so on, and then help them make quick and effective fire-fighting plan. Therefore, this paper proposes an early forest fire segmentation algorithm based on a deep learning model, named F-Unet, which mainly uses the architecture idea of Unet for reference. F-Unet consists of contraction path, feature fusion layer and expansion path. The contraction path is composed of the first 13 layers of VGG16, which is used to obtain feature maps with different scales. In order to improve the segmentation accuracy of the model, the feature fusion network proposed in this paper is added to the Unet architecture to fuse these feature maps with different scales. The expansion path is used for upsampling these feature maps to restore the size of the original input image and obtain the fire segmentation results. The experimental testing results on the FLAME dataset show that F-Unet can significantly improve the fire segmentation precision, and also prove that the proposed feature fusion network is effective to improve the performance of fire segmentation.

关键词： Deep learning image segmentation Fuses Fault detection Fires Forestry Safety

来源：评论

学校读者我要写书评

暂无评论

One-shot distributed algorithm for generalized eigenvalue problem

arXiv

引用

arXiv 2020年

作者： Lyu, Kexin He, Fan Huang, Xiaolin Yang, Jie Chen, Liming Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University MOE Key Laboratory of System Control and Information Processing 800 Dongchuan Road Shanghai200240 China Institute of Medical Robotics Shanghai Jiao Tong University 800 Dongchuan Road Shanghai200240 China Ecole Centrale de Lyon France

Nowadays, more and more datasets are stored in a distributed way for the sake of memory storage or data privacy. The generalized eigenvalue problem (GEP) plays a vital role in a large family of high-dimensional statistical models. However, the existing distributed method for eigenvalue decomposition cannot be applied in GEP for the divergence of the empirical covariance matrix. Here we propose a general distributed GEP framework with one-shot communication for GEP. If the symmetric data covariance has repeated eigenvalues, e.g., in canonical component analysis, we further modify the method for better convergence. The theoretical analysis on approximation error is conducted and the relation to the divergence of the data covariance, the eigenvalues of the empirical data covariance, and the number of local servers is analyzed. Numerical experiments also show the effectiveness of the proposed algorithms. Copyright © 2020, The Authors. All rights reserved.

关键词： Eigenvalues and eigenfunctions

来源：评论

学校读者我要写书评

暂无评论

SMTC-CL: Continuous Learning via Selective Multi-Task Coordination for Adaptive Signal Classification

引用

IEEE Transactions on Cognitive Communications and Networking 2024年

作者： Hao, Xiaoyang Yang, Shuyuan Liu, Ruoyu Feng, Zhixi Peng, Tongqing Huang, Bincheng Xidian University Key Laboratory of Intelligent Perception and Image Understanding Ministry of Education School of Artificial Intelligence Xi'an710071 China Avic Xi'an Flight Automatic Control Research Institute China MeiTuan China China Electronics Technology Group Corporation Institute of Information Science and Technology China

Continuously adaptive signal classification in complex electromagnetic environments is a desired property of realistic intelligent systems. However, the limitation of most existing signal processing tasks and methods lies in their assumption of fixed categories and invariant distributions. In contrast, humans naturally acquire knowledge continuously from data streams, enhancing their capabilities through cross-task knowledge integration and learning from the past. Inspired by this, we propose a continuous learning (CL) method based on selective multi-task coordination (SMTC) to deal with ever-changing task categories and distributions in practical electromagnetic environments (denoted as SMTC-CL). Firstly, we propose a signal task distribution comparison (TDC) method that leverages intra-domain density and inter-domain divergence. TDC selectively activates multiple old tasks to collaboration with the current incremental new task. Secondly, we propose the SMTC method. By revisiting potential collaborative tasks from the past using TDC and formulating different selection strategies based on the context, continuous learners can achieve mutual benefits between past and current tasks in SMTC learning. SMTC can partially alleviate catastrophic forgetting and enable continuous learners to adapt to the changing electromagnetic environment. Based on the signal automatic modulation classification (AMC) and real-world specific emitter identification (SEI) datasets, extensive experimental results validate the effectiveness of our proposed method on a variety of CL scenarios, including varying signal task categories, distributions, and their combinations. © 2015 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Research Progress of Encoder-Decoder Network in image Fusion

SSRN

引用

SSRN 2023年

作者： Zhou, Tao Zhang, Xiangxiang Cheng, Qianru Li, Qi Lu, Huiling School of Computer Science and Engineering North Minzu University Yinchuan750021 China Key Laboratory of Image and Graphics Intelligent Processing of State Ethnic Affairs Commission North Minzu University Yinchuan750021 China School of Science Ningxia Medical University Yinchuan750004 China

A B S T R A C Timage fusion is a research hotspot in the field of computer vision, which obtains complementary information from the different modal images and then presents it in a comprehensive, highly information-cohesive fused image. Encoder-decoder networks are an important technical method in the field of image fusion. This paper reviewed the research status and development of the Encoder-decoder networks in the field of image fusion from the following aspects: Firstly, the18 image fusion-related datasets and the 22 open-source codes of deep learning-based image fusion are summarized. Secondly, the model structure of encoder-decoder networks and the basic principles for image fusion based on the encoder-decoder network are described. Thirdly, encoder-decoder networks fusion models are summarized into 4 aspects: Convolutional Auto-encoder network(CAE Net), Sparse Auto-encoder network(SAE Net), U-shape Encoder-decoder network(U-net) and Variational Auto-encoder network(VAE Net),in U-net, there are 5 types: U-net incorporating residual connection mechanism (Res U-net), U-net incorporating dense connection mechanism (Dens U-net), U-net incorporating multiscale connection mechanism (MS U-net), U-net incorporating attention mechanism (Att U-net) and U-shaped Encoder-decoder networks(U-net) combined with Generative Adversarial Network(GAN U-net). Fourthly, the existing image fusion evaluation metrics are summarized from both subjective and objective aspects. Fifthly, the application of Encoder-decoder network in 5 kinds of image fusion, such as multi-exposure image fusion, multi-focus image fusion, infrared and visible image fusion, medical image fusion and remote sensing image fusion are summarized. Finally, there is a discussion of the main challenges faced by encoder-decoder networks in the image fusion domain and an outlook on future directions. This paper systematically analyzed the application of encoder-decoder networks in the image fusion areas, which has posi

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Multiple-atlas-based automatic sementation of hippocampus for lateralization in temporal lobe epilepsy

Multiple-atlas-based automatic sementation of hippocampus fo...

引用

IEEE International Symposium on Biomedical Imaging

作者： Alireza Akhondi-Asl Kourosh Jafari-Khouzani Hamid Soltanian-Zadeh Image Analysis Laboratory Department of Diagnostic Radiology Henry Ford Health System Detroit MI USA School of Cognitive Sciences Institute of Studies in Theoretical Physics and Mathematics Tehran Iran Control and Intelligent Processing Center of Excellence School of Electrical and Computer Engineering University of Tehran Tehran Iran

ISBN: (纸本)9781424441259

We introduce a 3D segmentation framework which uses principal shapes. The probabilistic energy function of the method is defined based on intensity, tissue type, and location information of the structures using a multiple atlas method. For intensity information, nonparametric probability density function is used which considers intensity relation of different structures. To find a local minimum of the energy function, a two-step optimization strategy is used. In the first step, shape parameters are optimized based on the analytic derivatives of the energy function. In the second step, shapes of the structures are fine-tuned using a level set method. The proposed method is shown to be superior to some popular methods in the literature using a dataset of 64 patients with mesial temporal lobe epilepsy. In addition, the method can be used for lateralization with accuracy close to that of manual segmentation.

关键词： Hippocampus Temporal lobe Epilepsy image segmentation Principal component analysis Shape control Probability density function Level set Skull Algorithm design and analysis

来源：评论

学校读者我要写书评

暂无评论

Differential Private Discrete Noise Adding Mechanism: Conditions, Properties and Optimization

arXiv

引用

arXiv 2022年

作者： Qin, Shuying He, Jianping Fang, Chongrong Lam, James The Dept. of Automation Shanghai Jiao Tong University The Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai Engineering Research Center of Intelligent Control and Management Shanghai China The Dept. of Mechanical Engineering The University of Hong Kong Pokfulam Road Hong Kong

Differential privacy is a standard framework to quantify the privacy loss in the data anonymization process. To preserve differential privacy, a random noise adding mechanism is widely adopted, where the trade-off between data privacy level and data utility is of great concern. The privacy and utility properties for the continuous noise adding mechanism have been well studied. However, the related works are insufficient for the discrete random mechanism on discretely distributed data, e.g., traffic data, health records. This paper focuses on the discrete random noise adding mechanisms. We study the basic differential privacy conditions and properties for the general discrete random mechanisms, as well as the trade-off between data privacy and data utility. Specifically, we derive a sufficient and necessary condition for discrete ∊-differential privacy and a sufficient condition for discrete (∊, δ)-differential privacy, with the numerical estimation of differential privacy parameters. These conditions can be applied to analyze the differential privacy properties for the discrete noise adding mechanisms with various kinds of noises. Then, with the differential privacy guarantees, we propose an optimal discrete ∊-differential private noise adding mechanism under the utility-maximization framework, where the utility is characterized by the similarity of the statistical properties between the mechanism's input and output. For this setup, we find that the class of the discrete noise probability distributions in the optimal mechanism is Staircase-shaped. Copyright © 2022, The Authors. All rights reserved.

关键词： Probability distributions

来源：评论

学校读者我要写书评

暂无评论

T-TIME: Test-Time Information Maximization Ensemble for Plug-and-Play BCIs

arXiv

引用

arXiv 2024年

作者： Li, Siyang Wang, Ziwei Luo, Hanbin Ding, Lieyun Wu, Dongrui Ministry of Education Key Laboratory of Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China Shenzhen Huazhong University of Science and Technology Research Institute Shenzhen China Henan Key Laboratory of Brain Science and Brain Computer Interface Technology School of Electrical and Information Engineering Zhengzhou University China School of Civil and Hydraulic Engineering Huazhong University of Science and Technology Wuhan430074 China

Objective: An electroencephalogram (EEG)-based brain-computer interface (BCI) enables direct communication between the human brain and a computer. Due to individual differences and non-stationarity of EEG signals, such BCIs usually require a subject-specific calibration session before each use, which is time-consuming and user-unfriendly. Transfer learning (TL) has been proposed to shorten or eliminate this calibration, but existing TL approaches mainly consider offline settings, where all unlabeled EEG trials from the new user are available. Methods: This paper proposes Test-Time Information Maximization Ensemble (T-TIME) to accommodate the most challenging online TL scenario, where unlabeled EEG data from the new user arrive in a stream, and immediate classification is performed. T-TIME initializes multiple classifiers from the aligned source data. When an unlabeled test EEG trial arrives, T-TIME first predicts its labels using ensemble learning, and then updates each classifier by conditional entropy minimization and adaptive marginal distribution regularization. Our code is publicized. Results: Extensive experiments on three public motor imagery based BCI datasets demonstrated that T-TIME outperformed about 20 classical and state-of-the-art TL approaches. Significance: To our knowledge, this is the first work on test time adaptation for calibration-free EEG-based BCIs, making plug-and-play BCIs possible. © 2024, CC BY.

关键词： Calibration

来源：评论

学校读者我要写书评

暂无评论

Prediction of piRNA-mRNA interactions based on an interactive inference network

Prediction of piRNA-mRNA interactions based on an interactiv...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Yajun Liu Ru Li Aimin Li Rong Fei Xie Guo Fang-Xiang Wu Shaanxi Key Laboratory for Network Computing and Security Technology Xi'an University of Technology Xi'an China Shaanxi Key Laboratory of Complex System Control and Intelligent Information Processing Xi'an University of Technology Xi'an China Department of Computer Science Department of Mechanical Engineering Division of Biomedical Engineering University of Saskatchewan Saskatoon Canada

As the largest class of small non-coding RNAs, piRNAs primarily present in the reproductive cells of mammals, which influence post-transcriptional processes of mRNAs in multiple ways. Effective methods for predicting piRNA and mRNA target relationships can help identify piRNA functions, investigate the possibility of piRNAs as biomarkers and therapeutic targets. In this study, we propose a computational approach for classifying the relationships of piRNA-mRNA pairs based on an interactive inference network (IIN). First, we gather piRNA-mRNA target data, collect sequence data by position alignment, and construct a benchmark dataset. Furthermore, a reliable negative set is constructed by positive-unlabeled learning. Finally, we view a piRNA and a mRNA sequence as a premise and hypothesis sentence, respectively, and IIN model is used to predict the relationship between them. The experiments demonstrate that our method effectively characterizes piRNA-mRNA interaction and could be beneficial for researchers to investigate piRNA functions.

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：