检索结果-内蒙古大学图书馆

2022 Wave Electronics and its Application in Information and Telecommunication Systems, WECONF 2022

作者： Pinchukov, V.V. Poroykov, A. Yu. Shmatko, E.V. Bogachev, A.D. Sivov, N. Yu. Named after V.A. Fabrikant National Research University Moscow Power Engineering Institute Department of Physics Moscow Russia

ISBN: (纸本)9781665462402

In this paper three methods for determining displacements on images are compared. Two of them are neural networks designed for Particle image Velocimetry images processing. The third method is classic cross-correlation approach. For estimation the accuracy of the algorithms synthetic data were used. Based on the obtained simulation results, it can be concluded that the use of neural networks is promising both for fluid dynamics and solid mechanics applications. © 2022 IEEE.

关键词： neural networks

来源：评论

学校读者我要写书评

暂无评论

AFHO-DL: Enhancing Energy Efficiency through Resource Allocation AI-enabled WSNs and IoT Integration 3

AFHO-DL: Enhancing Energy Efficiency through Resource Alloca...

引用

3rd International conference on Electrical, Electronics, Information and Communication Technologies, ICEEICT 2024

作者： Kalaiselvan, S.A. Manoranjini, J. Hemalatha, S. Kumar, M. Lenin Rajalakshmi Engineering College Department of Artificial Intelligence and Machine Learning Chennai India Vignan Institute of Technology and Sciences Department of Electronics Communication and Engineering Telangana India

ISBN: (纸本)9798350369083

In the realm of AI-enabled Wireless Sensor networks (WSNs) and Internet of Things (IoT) integration, efficient resource allocation is paramount for enhancing energy efficiency and optimizing data utilization. The dynamic nature of environments poses challenges to achieving optimal performance measures, necessitating sophisticated solutions. To address this, we propose the AFHO-DL model, which combines the Fire Hawk Optimizer (FHO), UNet architecture, and Multi-objective Jaya algorithm. The FHO algorithm, inspired by the foraging behavior of Fire Hawks, efficiently optimizes resource allocation in WSNs by iteratively updating node positions based on objective functions. Leveraging FHO, we fine-tune the parameters of the UNet architecture, a deep learning model widely used in image processing applications. The UNet architecture, consisting of encoder and decoder paths, extracts features and reconstructs images, enhancing adaptability to various data types. Furthermore, the Multi-objective Jaya algorithm is employed to refine solutions by striking a balance between exploration and exploitation in the solution space, further improving resource allocation strategies. Our experimental results demonstrate the effectiveness of the proposed AFHO-DL model in enhancing energy efficiency and optimizing data utilization in dynamic environments. Through post-processing and visualization, we evaluate the performance of the optimization algorithms, generating Pareto-optimal solutions that represent the best trade-offs between competing goals. The AFHO-DL model facilitates intelligent decision-making and resource allocation strategy optimization, ensuring smooth data transfer and improving network efficiency in AI-enabled WSNs coupled with IoT platforms. © 2024 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

A Verilog based Approach for Object Detection using CNN

A Verilog based Approach for Object Detection using CNN

引用

Global conference for Advancement in Technology (GCAT)

作者： Samson Swaraj Quadros S. Adityakrishna Kirti S. Pande Department of Electronics and Communication Engineering Amrita School of Engineering Amrita Vishwa Vidyapeetham Bengaluru India

ISBN: (数字)9798350376685

ISBN: (纸本)9798350376692

A Convolutional neural Network (CNN) are a class of artificial neural networks specifically designed to process data with a grid-like topology, such as images, making them well-suited for tasks like image recognition and classification, object detection, and speech recognition. However, their current software implementations leave much to be desired regarding energy efficiency, speed, performance and scalability. Hardware-based implementation of CNNs has several significant advantages over software, including faster data processing due to the parallel execution of hardware-based FPGAs and ASICs, which are necessary for real-time applications. They are more energy-efficient and have consistent, predictable performance. In this paper, we present the implementation of CNN on Verilog. We implemented a highly optimized CNN regression model architecture for object detection having an accuracy of 97%. The entire design was made on Verilog, allowing easy transferability to both FPGA and ASIC platforms. The proposed work is compared to the current standard for software implementations – Google Colab. The results obtained show considerable speed up and improved performance.

关键词： Scalability Object detection Speech recognition Software Energy efficiency Topology Convolutional neural networks Hardware design languages Field programmable gate arrays Standards

来源：评论

学校读者我要写书评

暂无评论

GDTNet: A Synergistic Dilated Transformer and CNN by Gate Attention for Abdominal Multi-organ Segmentation 30th

GDTNet: A Synergistic Dilated Transformer and CNN by Gate ...

引用

30th International conference on MultiMedia Modeling, MMM 2024

作者： Zhang, Can Wang, Zhiqiang Zhang, Yuan Li, Xuanya Hu, Kai Key Laboratory of Intelligent Computing and Information Processing of Ministry of Education Xiangtan University Xiangtan411105 China Key Laboratory of Medical Imaging and Artificial Intelligence of Hunan Province Xiangnan University Chenzhou423000 China Baidu Inc. Beijing100085 China

ISBN: (纸本)9783031533013

As one of the key problems in computer-aided medical image analysis, learning how to model global relationships and extract local details is crucial to improve the performance of abdominal multi-organ segmentation. While current techniques for Convolutional neural networks (CNNs) are quite mature, their limited receptive field makes it difficult to balance the ability to capture global relationships with local details, especially when stacked onto deeper networks. Thus, several recent works have proposed Vision Transformer based on a self-attentive mechanism and used it for abdominal multi-organ segmentation. However, Vision Transformer is computationally expensive by modeling long-range relationships on pairs of patches. To address these issues, we propose a novel multi-organ segmentation framework, named GDTNet, based on the synergy of CNN and Transformer for mining global relationships and local details. To achieve this goal, we innovatively design a Dilated Attention Module (DAM) that can efficiently capture global contextual features and construct global semantic information. Specifically, we employ a three-parallel branching structure to model the global semantic information of multiscale encoded features by Dilated Transformer, combined with global average pooling under the supervision of Gate Attention. In addition, we fuse each DAM with DAMs from all previous layers to further encode features between scales. Extensive experiments on the Synapse dataset show that our method outperforms ten other state-of-the-art segmentation methods, achieving accurate segmentation of multiple organs in the abdomen. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Computer aided analysis

来源：评论

学校读者我要写书评

暂无评论

27th International conference series on Climbing and Walking Robots and the Support Technologies for Mobile Machines, CLAWAR 2024

27th International Conference series on Climbing and Walking...

引用

27th International conference series on Climbing and Walking Robots and the Support Technologies for Mobile Machines, CLAWAR 2024

ISBN: (纸本)9783031707216

The proceedings contain 28 papers. The special focus in this conference is on Climbing and Walking Robots and the Support Technologies for Mobile Machines. The topics include: 27 Years of Climbing and Walking Robots – Are We There?;Torque Controlled or Intrinsically Compliant? DLR’s Perspective on Robust and Efficient Biped and Quadruped Locomotion;Efficient Stream-Based Active Learning Initialization for Legged Robots Based on a PCA/K-Means image Selection Approach;precision Vehicle Pose Estimation with Uncertainty-aware neural Network;HAPmamba: Linear-Time Sequence Modeling for Terrain Classification by Legged Robots;neural-Based Self-collision Checking for a Quadruped Robot;omnidirectional Climbing Robot for Maintenance Services on Hard to Reach Places of Ship Hulls;demonstration of a Micro Wall-Climbing Robot Moving on Metal Surfaces;linkage Length Optimization of a Climbing Inspection Robot Using an Area Overlap Method;Rotary Push-in Mechanism for Variable Outer Diameter PIGs with Multiple-Connected Using Pneumatic artificial Muscles;basic Study on a Peristaltic Motion-Type In-Pipe Inspection Robot Using a Hyper-Extension Unit for Improving Locomotion Speed;proposal of Operation Methods of the Square-Duct Cleaning Machine with Multistage Planetary Gear Mechanism;earth-Shaping with Heterogeneous Robotic Teams: From Sim to Real;mobile Victim Signs Monitoring Through Non-invasive Robotic System;Multi-UAV Coverage Path Planning for Agricultural applications;autonomous Landing Pad with a Closed Cover for a Medium-Sized Drone to Support Typical Research and Reconaissance Tasks in the Local Environment;concept of Pneumatic Soft Robot: Suction-Driven Locomotion;climbing Robot Inspired by Inchworms: Adaptable for Tubular and Flat Surfaces with Multi-plane Work Capability;high-Propulsive Trunk Flexion–Extension Mechanism Using Cheetah-Inspired S-Shaped Spine;self-organized Locomotion with Multiple Stepping Frequencies in an Insect-Like Robot Under Decentralized Adaptive

关键词：

来源：评论

学校读者我要写书评

暂无评论

Silicon Micro-Disk Resonator Crossbar Array for High-Speed and High-Density Photonic Convolution processing

arXiv

引用

arXiv 2025年

作者： Huang, Long Yao, Jianping Microwave Photonics Research Laboratory School of Electrical Engineering and Computer Science University of Ottawa OttawaONK1N 6N5 Canada

Advanced artificial intelligence (AI) algorithms, particularly those based on artificial neural networks, have garnered significant attention for their potential applications in areas such as image recognition and natural language processing. Notably, neural networks make heavy use of matrix-vector multiplication (MVM) operations, causing substantial computing burden on existing electronic computing systems. Optical computing has attracted considerable attention that can perform optical-domain MVM at an ultra-high speed. In this paper, we introduce a novel silicon photonic micro-disk resonator (MDR) crossbar signal processor designed to support matrix-vector multiplication (MVM) with both high processing speed and enhanced computational density. The key innovation of the proposed MDR crossbar processor is the placement of two MDRs at each crosspoint, enabling simultaneous routing and weighting functions. This design effectively doubles the computational density, improving overall performance. We fabricate a silicon photonic MDR crossbar processor, which is employed to perform convolutional tasks in a convolutional neural network (CNN). The experimental results demonstrate that the photonic processor achieves a classification accuracy of 96% on the MNIST dataset. Additionally, it is capable of scaling to a computational speed of up to 160 tera-operations per second (TOPS) and a computational density as high as 25.6 TOPS/mm2. Our approach holds significant promise for enabling highly efficient, scalable on-chip optical computing, with broad potential applications in AI and beyond. © 2025, CC BY-NC-ND.

关键词： Silicon photonics

来源：评论

学校读者我要写书评

暂无评论

image Captioning: Analyzing CNN-LSTM and Vision-GPT Models

Image Captioning: Analyzing CNN-LSTM and Vision-GPT Models

引用

International conference for Convergence of Technology (I2CT)

作者： Abburi Sai Karthik Maddala H S M Krishna Karthik Samudrala Yashwanth Anjali T Dept. of Computer Science and Engineering Amrita Vishwa Vidyapeetham Amritapuri India

ISBN: (数字)9798350394474

ISBN: (纸本)9798350394481

image captioning, which exists at the point of intersection of computer vision and natural language processing, is essential for enhancing image comprehension, allowing applications like content discovery, visual aid for the blind, and more. The hunt for more precise and reliable picture captioning models continues to be an important research goal as technology develops quickly. The two prominent image captioning techniques used in this study image Captioning Using LSTM+CNN and image Captioning Using VisionGPT2 are thoroughly compared. We examine these models' internal workings, assess their effectiveness, and offer insights into their advantages and disadvantages for diverse application *** neural networks (CNNs) for extracting visual features and long short-term memory (LSTM) networks for producing sequential language are combined in the LSTM+CNN model, a tried-and-true methodology. It has shown adept in creating insightful descriptions for a variety of photographs. On the other hand, VisionGPT2, a GPT-2 architectural extension, makes use of transformers and pretrained language models to provide cutting-edge outcomes in a range of natural language processing applications. We analyze the viability of each technique by taking into account elements like model complexity, training data needs, and deployment simplicity. This comprehensive comparison enlightens academics, programmers, and businesses on the ideal picture captioning solution for their particular requirements, fostering development in this area and its numerous uses.

关键词： Analytical models Visualization neural networks Training data Feature extraction Transformers Natural language processing

来源：评论

学校读者我要写书评

暂无评论

NN-Based In-Loop Filtering With Inputs Transformed

NN-Based In-Loop Filtering With Inputs Transformed

引用

IEEE International conference on image processing

作者： Du Liu Jacob Ström Mitra Damghanian Per Wennersten Ericsson Stockholm Sweden

ISBN: (数字)9798350349399

ISBN: (纸本)9798350349405

The state-of-the-art neural network-based (NN-based) in-loop filters for video coding are built on convolutional neural networks. The Joint Video Experts Team (JVET) activities investigate NN-based in-loop filters for two operation points, the high operation point (HOP) which provides highest possible gains at a high complexity and the low operation point (LOP) which is constrained on a low complexity. This paper focuses on the LOP network. We apply a DCT and reshaping to the inputs and an inverse DCT and inverse reshaping to the outputs of LOP. The spatial resolution inside the network is reduced by a factor of four while the final output still has the same number of pixels. The complexity in MAC/pixel (multiplyaccumulate operations per pixel) is therefore also reduced by a factor of four. This freed-up complexity is instead spent on increasing the number of backbone blocks and channels so the LOP complexity is matched. Our network has a complexity of $16.9 \mathrm{kMAC} /$ pixel and 0.2 M parameters (LOP: 17 kMAC/pixel, 0.05 M parameters). The BD-rate impact compared to the NNVC-7.1 anchor is reported to be −0.48% for RA and −0.17% for AI with the float model, and −0.44% for RA and −0.18% for AI with the integer model.

关键词： Video coding Filters image processing artificial neural networks Transforms Complexity theory Discrete cosine transforms

来源：评论

学校读者我要写书评

暂无评论

JFMNet: Joint Fusion Multi-networks for image Dehazing and Denoising in the Port Environment 2022

JFMNet: Joint Fusion Multi-Networks for Image Dehazing and D...

引用

14th International conference on Machine Learning and Computing, ICMLC 2022

作者： Lin, Guancheng Zheng, Yijie Xu, Zhihong Xia, Tianzhi Yuan, Peng School of Computer and Artificial Intelligence Wuhan University of Technology China School of Navigation Wuhan University of Technology China School of Naval Architecture Wuhan University of Technology China

ISBN: (纸本)9781450395700

The bad weather events, such as haze, in maritime traffic dramatically reduce the visibility, which can seriously affect the ship navigation especially in areas with intensive port traffic. Meanwhile, unwanted signals are inevitably introduced by the maritime imaging device during image capturing and transmission in hazy conditions. Therefore, the captured image is not only degraded by the haze, but also may contain unwanted noise. These low-quality images interfere with the subsequent image processing and increase the potential for maritime traffic accidents. It is therefore imperative to improve the image quality in hazy conditions. To reveal the information hidden in the haze while suppress noise, this paper proposes the joint fusion multi-networks (termed JFMNet) for image dehazing and denoising in the port environment. The multi-networks use the dehazing module (DHNet) and the denoising module (DNNet) to suppress the noise and haze. Then use the information fusion module (FNet) to integrate the results of the DNNet and DHNet with the information of the original input images to achieve the goal of dehazing and denoising while preserving the details. The modules in multi-networks are based on an encoder-decoder structure. Experiments on a number of challenging hazy images with noise are present to reveal the efficacy of this structure. Meanwhile, experiments also show our JFMNet's superiority over several state-of-the-arts in terms of dehaze quality and efficiency. © 2022 ACM.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Analyzing Chatbot Architectures Utilising Deep neural networks

Analyzing Chatbot Architectures Utilising Deep Neural Networ...

引用

Cybernetics, Cognition and Machine Learning applications (ICCCMLA), IEEE International conference on

作者： Kavita Patil Rohit Patil Vedanti Koyande Amaya Singh Thakur Kshitij Kadam Computer Science and Business System JSPM’s Rajarshi Shahu College of Engineering Pune India

ISBN: (数字)9798331505790

ISBN: (纸本)9798331505806

This survey paper examines the advancements and challenges in chatbot technology, focusing on deep neural networks (DNNs) and their application in natural language processing (NLP). It discusses various chatbot models, including Elizabot, Alicebot, Mitsuku, and Cleverbot, highlighting their evolution from rule-based systems to sophisticated AI conversational agents. The study introduces a specialized chatbot for website integration, emphasizing the importance of swift, accurate, and personalized interactions to enhance customer engagement. Additionally, the paper explores the integration of Large Language Models (LLMs) such as Gemini, GPT and BERT, fine-tuning with deep learning techniques to improve chatbot performance, and their potential to revolutionize customer interactions and business growth.

关键词： Surveys Deep learning Analytical models Large language models Focusing artificial neural networks Chatbots Cognition Cybernetics Business

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：