检索结果-内蒙古大学图书馆

CMANet: A CNN-Mamba aggregation network for face super-resolution

Pattern Recognition 2025年 168卷

作者： Wu, Ziyi Lu, Tao Zhang, Yanduo Chai, Xiaoyu Hubei Key Laboratory of Intelligent Robot Wuhan Institute of Technology Wuhan China Computer School Hubei University of Arts and Science Xiangyang China National Engineering Research Center for Multimedia Software Wuhan University Wuhan China

Currently, most face super-resolution (FSR) methods rely on convolutional neural networks (CNNs) or Transformers, both of which have inherent limitations: CNNs suffer from a limited receptive field, while Transformers introduce quadratic computational complexity due to self-attention. In recent years, Mamba has effectively addressed these challenges using the state space model (SSM), excelling in global modeling. However, due to the highly structured nature of face images, relying solely on global modeling is insufficient for high-quality reconstruction. It is crucial to preserve global structural information while enhancing local texture details. While Mamba excels in global modeling, it lacks the ability to effectively extract local details. To eliminate this issue, we propose the CNN-Mamba Aggregation Network (CMANet) for FSR, which consists of two branches: a CNN branch and a Mamba branch. The CNN branch is a traditional end-to-end FSR network, in which the proposed Multi-Scale Residual Block (MSRB) extracts multi-scale local details. The Mamba branch, the core of the network, is built upon a UNet architecture. Its central component is a carefully designed Dual Scan Mamba Block (DSMB), which combines the Dual Scan Mamba Module (DSMM) and Feature Refinement Feed-Forward Network (FR-FFN). The DSMM enhances the global modeling capabilities of the SSM in both plane and channel scanning directions, while the FR-FFN refines the expressiveness and adaptability of the features. Finally, CMANet reconstructs high-fidelity face images by aggregating features from both branches. Extensive experimental results show that our model significantly outperforms many advanced methods in both objective metrics and subjective image quality. © 2025 Elsevier Ltd

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Cancer Cell Detection and Classification from Digital Whole Slide Image 5th

Cancer Cell Detection and Classification from Digital Whole ...

引用

5th International Conference on Smart Technologies in Data science and Communication, SMART-DSC 2022

作者： Gavade, Anil B. Nerli, Rajendra B. Ghagane, Shridhar Gavade, Priyanka A. Bhagavatula, Venkata Siva Prasad Department of E&C KLS Gogte Institute of Technology Karnataka Belagavi India Karnataka Belagavi India Department of Biotechnology KAHER’s Dr. Prabhakar Kore Basic Science Research Center V. K. Institute of Dental Sciences Campus Karnataka Belagavi India Department of Computer Science and Engineering KLE Tech University Dr. M. S. Sheshgiri College of Engineering and Technology Karnataka Belagavi India Hyderabad Medtronic India

ISBN: (纸本)9789811968792

The World Health Organisation has identified cancer as one of the foremost causes of death globally which reports that nearly one in six deaths is due to cancer. Hence, an early and correct diagnosis is required to assist doctors in selecting the accurate and best treatment option for the patient. Pathological data have huge tumour information that can be used to diagnose cancer. Digitizing pathological data into images and its analysis using Deep learning applications will be a significant contribution to clinical testing. Due to advancements in technology, artificial intelligence (AI) and digital pathology can now be combined allowing for image-based diagnosis. This study uses residual networks (ResNet-50) and convolutional neural network (CNN), which is pre-trained on ImageNet dataset to train and categories lung histopathology images into non-cancerous, lung adenocarcinoma, and lung squamous cell carcinoma delivering an accuracy of 98.9%. Experimentation results show that the ResNet-50 model delivers finer classification results when compared to state-of-the-art methods. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Artificial intelligence Classification Convolutional neural network Deep learning Graphic processing unit Histopathological image Residual networks

来源：评论

学校读者我要写书评

暂无评论

Privacy-Aware Artificial Intelligence with Homomorphic Encryption using Machine Learning

Privacy-Aware Artificial Intelligence with Homomorphic Encry...

引用

Sustainable Computing and Smart Systems (ICSCSS), International Conference on

作者： B. Srinivasa Rao Saumitra Chattopadhyay Prashant Singh Bramah Hazela G. Sabarinathan Kalva Yamini Department of Computer Science and Engineering Bachupally Hyderabad Telangana India Department of Computer Science and Engineering Graphic Era Hill University Dehradun India Department of Computer Science and Engineering Amity School of Engineering and Technology Lucknow Amity University Uttar Pradesh India Department of Mathematics PSNA College of Engineering and Technology Dindigul Tamilnadu India Cyber Security in Computer Science and Engineering Sri Ramachandra Institute of Higher Education and Research Chennai Tamilnadu India

Along with the expansion of machine learning (ML) applications, the amount of data required to create predictions increases. Big-data ML has always been limited by off-chip memory capacity and computational speed. Considerably, privacy is one of the limitations of big data, which can be solved by homomorphic encryption (HE). Due to the combination of HE and ML, the multi-party privacy-protected ML suggested in this research may assist numerous users in doing artificial intelligence (AI) without disclosing private data. The technique may train common models in situations of data abuse, particularly in private data protection. The model trained using the ML technique named Artificial Neural Network (ANN) has a similar impact to the model developed using all data on a single computer, according to experiments using the algorithm. The gradient data is simply transmitted by all parties, and homomorphic procedures in the main computing system combine the gradient data. Besides, the optimal key is selected using the significance of the Lion Algorithm (LA). After homomorphic procedures, the learning model is modified depending on the new gradient data.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Synergistic effects of nano-enhanced waste transformer oil and hydrogen premixing on CI engine performance and emissions

引用

International Journal of Hydrogen Energy 2024年 56卷 484-497页

作者： Sathish, T. Karthikeyan, Sivakumar Sathyamurthy, Ravishankar Rajaram, K. Sandeep Kumar, S. Suresh Kumar, P. Giri, Jayant Amesho, Kassian T.T. Department of Mechanical Engineering Saveetha School of Engineering SIMATS Tamil Nadu Chennai India Department of Mechanical Engineering SRM Valliammai Engineering College Tamil Nadu Chennai603203 India Department of Mechanical Engineering KPR Institute of Engineering and Technology Tamil Nadu Coimbatore641407 India Department of Mechanical Engineering University Centre for Research and Development Chandigarh University Punjab Mohali India Department of Physics Gandhigram Rural Institute Deemed to be University Tamil Nadu Dindigul624302 India Department of Computer Science and Engineering Koneru Lakshmaiah Education Foundation Vaddeswaram Andhra Pradesh 522502 India Department of Mechanical Engineering R. V. R. and J.C. College of Engineering Andhra Pradesh Guntur522019 India Institute of Environmental Engineering National Sun Yat-Sen University Kaohsiung804 Taiwan Center for Emerging Contaminants Research National Sun Yat-Sen University Kaohsiung804 Taiwan Tshwane School for Business and Society Faculty of Management of Sciences Tshwane University of Technology Pretoria South Africa The International University of Management Centre for Environmental Studies Main Campus Dorado Park Ext 1 Windhoek Namibia Regent Business School Durban4001 South Africa Destinies Biomass Energy and Farming Pty Ltd SwakopmundP.O. Box 7387 Namibia Department of Mechanical Engineering Yeshwantrao Chavan College of Engineering Maharashtra Nagpur India

Waste materials possess untapped potential when harnessed effectively. One such resource is waste transformer oil, derived from electrical transformers, which can be converted into a valuable fuel source for compression-ignition (CI) engines using pyrolysis techniques. In this study, we explored the utilization of a 50 % blend of waste transformer oil with diesel, augmented with Titanium dioxide nanoparticles (referred to as "nano fuel of waste transformer oil"), in an unmodified CI engine. To further enhance the performance of this nano fuel, we introduced hydrogen (at flow rates of 2.5 LPM, 5.0 LPM, and 7.5 LPM) into the inlet section during the suction phase, promoting premixing within the cylinder. Hydrogen premixing significantly improved combustion dynamics. Notably, when 7.5 LPM of hydrogen was introduced alongside waste transformer oil-derived nano fuel, a remarkable 37.30 % increase in brake thermal efficiency was achieved. Moreover, this configuration exhibited a 15.33 % reduction in brake-specific fuel consumption (BSFC), a 27.09 % decrease in carbon monoxide (CO) emissions, a 37.50 % decrease in unburned hydrocarbon (UHC) emissions, and a 28.53 % reduction in smoke opacity compared to traditional diesel operation. These results highlight the potential for enhanced CI engine performance and reduced emissions through the strategic combination of waste transformer oil-based nano-fuel and hydrogen premixing. Consequently, we recommend further exploration of hydrogen premixing in conjunction with nano-enhanced waste transformer oil for superior CI engine performance. © 2023 Hydrogen Energy Publications LLC

关键词： Insulating oil

来源：评论

学校读者我要写书评

暂无评论

Reliable and Efficient Routing for Water Quality Monitoring in Underwater WSN

Reliable and Efficient Routing for Water Quality Monitoring ...

引用

computer, Communication and Control (IC4), International Conference on

作者： J. Bennilo Fernandes B. Gopi M. Ahsan Shariff Mohammed Rehaman Pasha S. Srinivasan S.P. Maniraj Department of Electronics and Communication Engineering Kalasalingam Academy of Research and Education Virudhunagar Tamil Nadu India Department of Electronics and Communication Engineering Muthayammal Engineering College (Autonomous) Rasipuram Tamil Nadu India Department of Biomedical Engineering Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Saveetha University Chennai Tamil Nadu India Department of Computer Science and Engineering Malla Reddy Institute of Engineering and Technology Hyderabad Telangana India M.Tech Computer Science and Engineering Sri Sai Ram Engineering College Chennai Tamil Nadu India Department of Computer Science and Engineering SRM Institute of Science and Technology Chennai Tamil Nadu India

Monitoring water quality has become crucial to provide clean and safe water. The traditional monitoring procedures are useless because they take a lot of function, require a long time, and cannot give results instantly to promote proactive water contamination reactions. This paper introduces a Reliable and efficient routing for Water Quality Monitoring (REWQ) in Underwater Wireless Sensor Networks (UWSNs). This mechanism utilizes the Cuckoo search (CS) algorithm to select the next hop to transmit the data from the sender to the surface region sink. The CS algorithm computes the fitness function by node energy usage, depth value and loss ratio. The proposed CS algorithm aims to improve reliability and successful data forwarding. This model is implemented, and its performances are evaluated by network simulator-NS-3. From the simulation results, the REWQ mechanism minimizes the average deliverance time and improves the residual energy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Remote Entangling Gates for Spin Qubits in Quantum Dots using a Charge-Sensitive Superconducting Coupler

arXiv

引用

arXiv 2024年

作者： Kang, Harry Hanlim Rosen, Ilan T. Hays, Max Grover, Jeffrey A. Oliver, William D. Research Laboratory of Electronics Massachusetts Institute of Technology CambridgeMA02139 United States Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology CambridgeMA02139 United States Department of Physics Massachusetts Institute of Technology CambridgeMA02139 United States

We propose a method to realize microwave-activated CZ gates between two remote spin qubits in quantum dots using a charge-sensitive superconducting coupler. The qubits are longitudinally coupled to the coupler, so that the transition frequency of the coupler depends on the logical qubit states;a capacitive network model using first-quantized charge operators is developed to illustrate this. Driving the coupler transition then implements a conditional phase shift on the qubits. Two pulsing schemes are investigated: a rapid, off-resonant pulse with constant amplitude, and a pulse with envelope engineering that incorporates dynamical decoupling to mitigate charge noise. We develop non-Markovian time-domain simulations to accurately model gate performance in the presence of 1/fβ charge noise. Simulation results indicate that a CZ gate fidelity exceeding 90% is possible with realistic parameters and noise models. Copyright © 2024, The Authors. All rights reserved.

关键词： Nanocrystals

来源：评论

学校读者我要写书评

暂无评论

Developing Intelligent AI-Driven Systems for Automated Data science

Developing Intelligent AI-Driven Systems for Automated Data ...

引用

Emerging research in Computational science (ICERCS), International Conference on

作者： Komal M Gajendra Shrimal Akhilendra Pratap Singh Solomon Jebaraj Lovish Dhingra Reena R Department of Information Technology Vishwakarma Institute of Information Technology Pune India Department of Computer Science & Engineering Vivekananda Global University Jaipur India Maharishi School of Engineering and Technology Maharishi University of Information Technology Uttar Pradesh India Department of Computer Science and Information Technology Jain (Deemed to be University) Bangalore Karnataka India Centre of Interdisciplinary Research in Business and Technology Chitkara University Institute of Engineering and Technology Chitkara University Punjab India Department of Computer Science and Engineering Prince Shri Venkateshwara Padmavathy Engineering College Chennai India

this article proposes the idea of applying smart synthetic intelligence (AI) structures to the automation of statistics technology. The AI structures are designed to study from records inputs to as it should be identify and procedure patterns, tendencies, and relationships within the statistics. Furthermore, those AI structures can be used to create computerized records technology fashions and practice them to datasets to be able to automate the records analysis procedure. The proposed AI gadget architecture is composed of numerous components, which might be accountable for extracting and extracting records, analyzing the statistics to decide styles and trends, and eventually, building models. The AI device structure contains superior strategies inclusive of deep gaining knowledge of, generative antagonistic networks, and reinforcement gaining knowledge of, so as to learn from a wide variety of records sources. Sooner or later, this newsletter discusses the capability advantages of the use of AI-pushed facts science automation and provides a few viable packages.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Domain adaptive FL for edge-enabled privacy-preserving MRI analysis

引用

Alexandria engineering Journal 2025年 128卷 324-339页

作者： Iqbal, Saeed Choudhry, Imran Arshad Ullah, Inam Kaur, Kuljeet Choi, Bong Jun Hassan, Mohammad Mehedi College of Mechatronics and Control Engineering Shenzhen University Shenzhen518060 China Department of Computer Science Faculty of Information Technology & Computer Science University of Central Punjab Lahore Pakistan Department of Computer Engineering Gachon University Seongnam13120 Korea Republic of Electrical Engineering Department École de technologie supérieure MontréalQCH3C1K3 Canada Centre for Research Impact & Outcome Chitkara University Institute of Engineering and Technology Chitkara University Punjab Rajpura140401 India School of Computer Science and Engineering Soongsil University Seoul Korea Republic of Department of Information Systems College of Computer and Information Sciences King Saud University Riyadh11543 Saudi Arabia

Data heterogeneity, privacy leakage challenges, the ineffectiveness of conventional collaborative learning techniques, and unresolved managing non-IID data distributions are some of the major obstacles to implementing artificial intelligence (AI) in healthcare. Federated learning (FL) frameworks frequently have trouble distinguishing between privacy protection and model accuracy, especially when used for delicate medical imaging applications. This study presents a novel framework that synergizes federated learning (FL) with edge computing to address these issues while safeguarding patient privacy. Our proposed Domain Adaptive Federated (DAD) learning approach effectively mitigates both inter-client and intra-client data heterogeneity, enabling collaborative model training across diverse medical imaging modalities (MRI, CT, PET) through cross-domain adaptation. Experimental evaluations on MRI brain segmentation datasets demonstrate the superior performance of DAD compared to traditional FL methods, as evidenced by significant improvements in F1-score (96.3), sensitivity (96.0), specificity (97.1), and AUC (96.7). This enhanced accuracy and robustness in handling heterogeneous and privacy-sensitive data render DAD an ideal candidate for privacy-preserving AI in consumer healthcare. By pioneering innovative strategies for collaborative model training and data privacy, this research contributes to the emerging field of edge intelligence, paving the way for improved patient outcomes while adhering to stringent confidentiality and ethical mandates. © 2025

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Smooth Tensor Qatar Riyal Decomposition for Dynamic MRI Reconstruction

引用

IEEE Journal of Biomedical and Health Informatics 2023年第6期29卷 3842-3852页

作者： Tingting Xu Yongyong Chen Haijin Zeng Guokai Zhang Jingyong Su School of Computer Science and Technology Harbin Institute of Technology Shenzhen China Guangdong Provincial Key Laboratory of Novel Security Intelligence Technologies Shenzhen China Image Processing and Interpretation Imec research group Ghent University Ghent Belgium School of Optical-Electrical and Computer Engineering University of Shanghai for Science and Technology Shanghai China

Dynamic magnetic resonance imaging (dMRI) speed and imaging quality have always been a crucial issue in medical imaging research. Most existing methods characterize the tensor rank-based minimization to reconstruct dMRI from sampling $\bf k$ - $t$ space data. However, (1) these approaches that unfold the tensor along each dimension destroy the inherent structure of dMR images. (2) they focus on preserving global information only, while ignoring the local details reconstruction such as the spatial piece-wise smoothness and sharp boundaries. To overcome these obstacles, we suggest a novel low-rank tensor decomposition approach by integrating tensor Qatar Riyal (QR) decomposition, low-rank tensor nuclear norm, and asymmetric total variation to reconstruct dMRI, named TQRTV. Specifically, while preserving the tensor inherent structure by utilizing tensor nuclear norm minimization to approximate tensor rank, QR decomposition reduces the dimensions in the low-rank constraint term, thereby improving the reconstruction performance. TQRTV further exploits the asymmetric total variation regularizer to capture local details. Numerical experiments demonstrate that the proposed reconstruction approach is superior to the existing ones.

关键词： Tensors Image reconstruction Matrix decomposition Magnetic resonance imaging Imaging Transforms Discrete Fourier transforms

来源：评论

学校读者我要写书评

暂无评论

Is Complexity Required for Neural Network Pruning? A Case Study on Global Magnitude Pruning

Is Complexity Required for Neural Network Pruning? A Case St...

引用

Artificial Intelligence (CAI), IEEE Conference on

作者： Manas Gupta Efe Camci Vishandi Rudy Keneta Abhishek Vaidyanathan Ritwik Kanodia Ashish James Chuan-Sheng Foo Min Wu Jie Lin Institute for Infocomm Research (I2R) Agency for Science Technology and Research (A*STAR) Singapore School of Computing (SoC) National University of Singapore (NUS) Singapore School of Computer Science and Engineering (SCSE) Nanyang Technological University (NTU) Singapore Centre for Frontier AI Research (CFAR) Agency for Science Technology and Research (A*STAR) Singapore

ISBN: (数字)9798350354096

ISBN: (纸本)9798350354102

Pruning neural networks has become popular in the last decade when it was shown that a large number of weights can be safely removed from modern neural networks without compromising accuracy. Numerous pruning methods have been proposed since, each claiming to be better than prior art, however, at the cost of increasingly complex pruning methodologies. These methodologies include utilizing importance scores, getting feedback through back-propagation or having heuristics-based pruning rules amongst others. In this work, we question whether this pattern of introducing complexity is really necessary to achieve better pruning results. We benchmark these SOTA techniques against a simple pruning baseline, namely, Global Magnitude Pruning (Global MP), that ranks weights in order of their magnitudes and prunes the smallest ones. Surprisingly, we find that vanilla Global MP performs very well against the SOTA techniques. When considering sparsity-accuracy trade-off, Global MP performs better than all SOTA techniques at all sparsity ratios. When considering FLOPs-accuracy trade-off, some SOTA techniques outperform Global MP at lower sparsity ratios, however, Global MP starts performing well at high sparsity ratios and performs very well at extremely high sparsity ratios. Moreover, we find that a common issue that many pruning algorithms run into at high sparsity rates, namely, layer-collapse, can be easily fixed in Global MP. We explore why layer collapse occurs in networks and how it can be mitigated in Global MP by utilizing a technique called Minimum Threshold. We showcase the above findings on various models (WRN-28-8, ResNet-32, ResNet-50, MobileNet-V1 and FastGRNN) and multiple datasets (CIFAR-10, ImageNet and HAR-2).

关键词： Costs Art Accuracy Computational modeling Artificial neural networks Benchmark testing Network architecture

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：