检索结果-内蒙古大学图书馆

Vesicle fusion and release in neurons under dynamic mechanical equilibrium

iScience 2024年第5期27卷 109793页

作者： Liu, Wenhao Gao, Tianyu Li, Na Shao, Shuai Liu, Bo Cancer Hospital of Dalian University of Technology Shenyang 110042 China Faculty of Medicine Liaoning Key Lab of Integrated Circuit and Biomedical Electronic System Dalian University of Technology Dalian 116024 China

Vesicular fusion plays a pivotal role in cellular processes, involving stages like vesicle trafficking, fusion pore formation, content release, and membrane integration or separation. This dynamic process is regulated by a complex interplay of protein assemblies, osmotic forces, and membrane tension, which together maintain a mechanical equilibrium within the cell. Changes in cellular mechanics or external pressures prompt adjustments in this equilibrium, highlighting the system's adaptability. This review delves into the synergy between intracellular proteins, structural components, and external forces in facilitating vesicular fusion and release. It also explores how cells respond to mechanical stress, maintaining equilibrium and offering insights into vesicle fusion mechanisms and the development of neurological disorders. © 2024 The Authors

关键词： Biophysics Neuroscience

来源：评论

学校读者我要写书评

暂无评论

Multiple Instance Learning for Lymph Node Metastasis Prediction from Cervical Cancer MRI

Multiple Instance Learning for Lymph Node Metastasis Predict...

引用

IEEE International Symposium on biomedical Imaging

作者： Shan Jin Hongming Xu Yue Dong Xinyu Hao Fengying Qin Ranran Wang Fengyu Cong School of Biomedical Engineering Faculty of Medicine Dalian University of Technology China Key Laboratory of Integrated Circuit and Biomedical Electronic System Liaoning Province China Department of Radiology Liaoning Cancer Hospital and Institute China Faculty of Information Technology University of Jyvaskyla Jyvaskyla Finland

Lymph node metastasis (LNM) is an important prognostic factor for recurrence and overall survival of cancer patients. The current LNM diagnosis is based on histopathologic examination after surgical lymphadenectomy, but an accurate and noninvasive method for LNM diagnosis is essential in selecting reasonable surgical operations and treatment plans. This paper presents an attention based multiple instance learning (MIL) model to diagnose LNM from cervical cancer multimodal MRI. The proposed MIL model adopts convolutional neural network (CNN) to extract features from multimodal MRI and attention-based pooling to make patient-level LNM status prediction. By incorporating the MIL and attention mechanism, the top rank MRI slice with informative regions in each LNM positive patient is visualized to provide the interpretability for LNM diagnosis. Experiments evaluated on a cohort of 241 cervical cancer patients show improvements in LNM status prediction compared with existing comparative models, which indicates the advantages of our designed model.

关键词：

来源：评论

学校读者我要写书评

暂无评论

High-Speed Transcendental Function Operation Unit Design 9

High-Speed Transcendental Function Operation Unit Design

引用

9th IEEE International Conference on Cyber Security and Cloud Computing and 8th IEEE International Conference on Edge Computing and Scalable Cloud, CSCloud-EdgeCom 2022

作者： Ze, Tian Feihu, Fan Jun, Zhang Xianglong, Ren Yang, Wei Xi'an XiangTeng Micro-Electronic Technology Co. Ltd Xi'an710068 China Key Laboratory of Aviation Science and Technology on Integrated Circuit and Micro-System Desigen Xi'an710068 China Northeastern Polytechnic University School of Computer Science Xi'an China

ISBN: (数字)9781665480666

ISBN: (纸本)9781665480666

With the rapid popularization and rapid development of embedded devices such as mobile terminals, people have become increasingly demanding for high-resolution picture rendering effects and fast computing capabilities of various processors. The enhancement of these capabilities of the processor depends on the computer's floating-point computing capabilities, which makes people to the GPU's rendering quality and speed, mobile devices, digital signal processing (DSP), and other computing units' floating-point computing capabilities Higher requirements have been raised [1]. The transcendental function arithmetic unit based on single-precision floating-point numbers is a very important component in these devices. For example, the perspective division, vector normalization, and coordinate transformation required in the GPU all require transcendental function operations [2] (For example, find the inverse, square root, trigonometric function, exponent, logarithm, etc.). In addition to the algebraic operations implemented in the field of mathematics, it is difficult to directly use a computer to calculate some complex functions. It can only complete complex functions through a limited number of simple operations, but for transcendental functions, in other words, the relationship between their input variables and output variables cannot be expressed by a finite number of basic operations such as addition, subtraction, multiplication, division, and power. Therefore, in a computer, the operation of a transcendental function cannot be directly implemented using a limited number of basic function operation units. A special operation unit is required to complete the operation of the transcendental function. © 2022 IEEE.

关键词： Polynomial approximation

来源：评论

学校读者我要写书评

暂无评论

Design of Shared Register File of GPU Unified Shader Array 9

Design of Shared Register File of GPU Unified Shader Array

引用

9th IEEE International Conference on Cyber Security and Cloud Computing and 8th IEEE International Conference on Edge Computing and Scalable Cloud, CSCloud-EdgeCom 2022

作者： Ze, Tian Xianglong, Ren Jun, Zhang Feihu, Fan Yue, Chen Xi'an XiangTeng Micro-Electronic Technology Co. Ltd Xi'an710068 China Key Laboratory of Aviation Science and Technology on Integrated Circuit and Micro-System Desigen Xi'an710068 China Northeastern Polytechnic University School of Computer Science Xi'an China

ISBN: (数字)9781665480666

ISBN: (纸本)9781665480666

Unified Shader Array is the computation core of the Unified Shader Array GPU, and the shader cores are the basic shader unit. In order to support the large-scale and thread-level parallelism of GPU, a large number of streaming processors composed of shader cores are set in GPU. The streaming processors can enable GPU to implement thread-level parallelism in the way of SIMD or SIMT. At the same time, GPU deploys large-scale register files for the streaming processor to reduce the cost of context switching, so the management of register files is of great significance to GPU performance. In the traditional GPU, each shader core has a unique register file, which simplifies the hardware management of the register. However, it will cause the limitation of the shader core's scale and insufficient register utilization. Based on the domestic special GPU, this paper analyzes the traditional GPU register scheduling strategy and carries out the design of the unified shader array shared register file based on pure hardware to reduce the cost of shader cores. This paper introduces its implementation in detail, including dynamic allocation and collection management of register files, allocation of multiple registers for thread bundles, detecting and handling of bank conflicts, and related register allocation and release collection instructions. Finally, we complete the RTL code implementation and analyze the experimental results. The results show that the design of this paper has reached the expected design goals in terms of simulation and logic synthesis. © 2022 IEEE.

关键词： Graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

A High Energy Efficiency and Low Resource Consumption FPGA Accelerator for Convolutional Neural Network 7

A High Energy Efficiency and Low Resource Consumption FPGA A...

引用

7th International Conference on Computer and Communications, ICCC 2021

作者： Zhang, Jianwei Cai, Houjia Li, Jinghu Dalian University of Technology Liaoning Key Lab of Integrated Circuit and Biomedical Electronic System School of Biomedical Engineering Dalian China Fujian Agriculture and Forestry University College of Computer and Information Sciences Fuzhou China

ISBN: (纸本)9781665409506

With the rapid development of convolutional neural networks (CNN), the design of hardware accelerators for CNN calculations has become a major focus of current research. However, power consumption and resource consumption are two main limits for the application of CNN hardware accelerators. In response to these problems, a FPGA-based implementation of a CNN hardware accelerator with high energy efficiency and low resource consumption is proposed. The implementation includes zero padding, convolution, batch normalization, ReLU6, and MaxPooling. The computing performance of the network is improved by using pipeline design and layer fusion. The LeNet-5 network is to test the implementation on the Xilinx Virtex VC-707 FPGA board. The experimental result shows that when inferring the MNIST test set, the energy efficiency is 843.91 times and 228.35 times that of the CPU and GPU, respectively. © 2021 IEEE.

关键词： Field programmable gate arrays (FPGA)

来源：评论

学校读者我要写书评

暂无评论

A High-Performance Hardware Accelerator for Sparse Convolutional Neural Network on FPGA

A High-Performance Hardware Accelerator for Sparse Convoluti...

引用

International Conference on Computer and Communications (ICCC)

作者： Jianwei Zhang Fen Xu Jinghu Li Liaoning Key Lab of Integrated Circuit and Biomedical Electronic System School of Biomedical Engineering & School of Artificial Intelligence Dalian University of Technology Dalian China Liaoning Key Lab of Integrated Circuit and Biomedical Electronic System School of Biomedical Engineering Dalian University of Technology Dalian China College of Computer and Information Sciences Fujian Agriculture and Forestry University Fuzhou China

Recently convolutional neural networks (CNNs) have developed rapidly and shown excellent performance in a wide range of applications. The design of hardware accelerators for convolutional neural networks has attracted great attention. With the number of layers in a CNN network deepens, the amount of model computation increases. The high computation and memory requirements can limit the deployment of CNNs on resource-limited hardware accelerators such as FPGAs, and it is urgent to use pruning technology to compress network models for easy deployment on FPGAs. In this paper, we use a sparse way based on the combination of convolutional kernel pruning and weight pruning to compress CNN models, together with 16-bit fixed-point quantization, skip zero-value processing, and independent multiply-and-accumulate (MAC) modules to implement a FPGA-based high-performance hardware accelerator design for sparse convolutional neural networks. The LeNet-5 network is implemented as an example on the Xilinx Virtex VC-707 FPGA board, and the experimental result from inference of 10,000 images of the MNIST test set shows that when the network sparsity is 87.46%, the inference time is 82.43% less than the CPU, the energy efficiency ratio is 264.85 times that implemented on the CPU.

关键词： Quantization (signal) Computational modeling Energy efficiency Convolutional neural networks Resource management Kernel Field programmable gate arrays

来源：评论

学校读者我要写书评

暂无评论

Enhancing Performance of Linked Independent Component Analysis: Investigating the Influence of Subjects and Modalities

Enhancing Performance of Linked Independent Component Analys...

引用

Computers, Information Processing and Advanced Education (CIPAE), 2020 International Conference on

作者： Huashuai Xu Huanjie Li Tommi Kärkkäinen Fengyu Cong Faculty of Information Technology University of Jyväskylä Jyväskylä Finland School of Biomedical Engineering Dalian University of Technology Dalian China School of Artificial Intelligence Faculty of Electronic Information and Electrical Engineering Key Laboratory of Integrated Circuit and Biomedical Electronic System School of Biomedical Engineering Faculty of Electronic Information and Electrical Engineering Dalian University of Technology Dalian China

In recent years, neuroimaging studies have increasingly been acquiring multiple modalities of data. The benefit of integrating multiple modalities through fusion lies in its ability to combine the unique strengths of each modality when analyzed collectively, as opposed to examining each one individually. In 2011, Adrian R. Groves proposed the Linked independent component analysis (LICA) method, which simultaneously models and discovers common features across multiple modalities. LICA has emerged as a powerful technique for analyzing multivariate data, particularly in neuroimaging and biomedical signal processing. The performance of LICA can be affected by the number of subjects and modalities. However, the detailed influence of the number of subjects and modalities on its performance remains an open question. In this study, we test the effects of the number of subjects and modalities on the performance of LICA using both simulated multimodal MRI data and the real multimodal MRI datasets from Autism Brain Imaging Data Exchange II (ABIDE II). Simulated data were utilized to evaluate the influence of subjects and modalities' variabilities. Real multi-site MRI data were used to demonstrate the advantages of multimodal fusion in identifying site-related components and removing site effects. Based on the simulation results, we found that increasing the number of modalities and subjects can improve the results when LICA can not recover the spatial maps or subject courses well. The correlation among subject courses from various modalities, the number of modalities, and the choice of components for decomposition all affect the linking performance of LICA. Our results from real-world datasets also demonstrated the advantages of multimodal fusion by LICA: 1) identify more site-related components; 2) remove more site effects.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Oxygen Vacancy Conductive Behavior on the Grain Boundaries of the Functional Layer of Hfo2-Based Rram

SSRN

引用

SSRN 2024年

作者： Qu, Dehao Liu, Junzhe Cui, Pengfei Qin, Fuwen Su, Yan Wang, Dejun Ministry of Education School of Control Science and Engineering Faculty of Electronic Information and Electrical Engineering Dalian University of Technology Dalian116024 China State Key Laboratory of Materials Modification by Laser Ion and Electron Beams Ministry of Education Dalian University of Technology Dalian116024 China Liaoning Key Laboratory of Integrated Circuit and Biomedical Electronic System Dalian University of Technology Dalian116024 China

In the study of HfO2-based resistive random access memory (RRAM), the prevailing mechanism of conducting filaments (CFs) is thought to primarily occur at grain boundaries (GBs). The current experimental study on the impact of oxygen vacancies (VOs) behavior at GBs on devices has limitations. Thus, this research utilized first principles calculations to delve into the structural modifications, energetics, density of states (DOS), and partial charge densities of VO configurations at GBs. The optimized structure analysis demonstrates that VOs exhibit an attraction towards neighboring oxygen atoms. Calculations of formation energies suggest that VOs are more likely to occur at GBs rather than in the bulk phase. The DOS and charge densities indicate that the GB significantly reduces the band gap. Furthermore, VO configurations lead to three distinct combinations of defect states, each corresponding to connected or localized charge densities. The VO12 and VO45 configurations enhance conductivity by forming charge channels, indicating that they may be the origin of the CFs at GBs. © 2024, The Authors. All rights reserved.

关键词： Oxygen vacancies

来源：评论

学校读者我要写书评

暂无评论

2D/3D Shape Model Registration with X-ray Images for Patient-Specific Spine Geometry Reconstruction 16th

2D/3D Shape Model Registration with X-ray Images for Patient...

引用

16th International Conference on Intelligent Robotics and Applications, ICIRA 2023

作者： Jiang, Yanxin Zhai, Haoyu Li, Kang Li, Lei Tao, Hairong Shao, Moyu Cheng, Xiaomin Wang, Hongkai School of Biomedical Engineering Faculty of Medicine Dalian University of Technology Dalian116024 China West China Biomedical Big Data Center West China Hospital Sichuan University Sichuan Chengdu610041 China Department of Vascular Surgery The Second Affiliated Hospital of Dalian Medical University Liaoning Dalian China Shanghai Key Laboratory of Orthopaedic Implants Department of Orthopaedic Surgery Shanghai Ninth People’s Hospital Shanghai Jiao Tong University School of Medicine Shanghai200011 China Jiangsu Yunqianbai Digital Technology Co. Ltd. Xuzhou221000 China Liaoning Key Laboratory of Integrated Circuit and Biomedical Electronic System Dalian116024 China

ISBN: (纸本)9789819964796

The use of preoperative CT and intraoperative fluoroscopic-guided surgical robotic assistance for spinal disease treatment has gained significant attention among surgeons. However, the intraoperative robotic systems lack guidance based on three-dimensional anatomical structures, rendering them unusable when there is a mismatch between preoperative CT and intraoperative fluoroscopy. Additionally, the continuous X-ray imaging during fluoroscopy exposes the patient to significant radiation risks. This paper proposes a 2D/3D registration method to reconstruct 3D patient spine geometry from frontal-view X-ray images. Our aim is to combine this method with intraoperative robotic assistance for spine procedures, enabling three-dimensional spatial navigation for the robotic system and reducing patient radiation exposure. By utilizing 45 frontal-view X-ray images of patients with scoliosis, we trained a 3D trunk anatomy SSM specific to scoliosis. For the registration, a 2D registration-back projection strategy was employed, which first generates the simulated X-ray projection image of the 3D SSM, then non-rigidly registers the projection image with the patient X-ray image, and finally back-project the 2D registration into 3D to guide the shape deformation of the 3D SSM. This approach was evaluated using X-ray images of 10 scoliosis patients. The results demonstrate the ability to reconstruct the anatomical structure and curvature of the spine in X-ray images. The registration accuracy of the SSM was assessed by visualizing the alignment between the registered model and X-ray images, as well as calculating the average distance between the vertebral disc center points of the model and expert-annotated anatomical landmarks in the patient’s X-rays. The average distance between the registered model’s vertebral disc centers and the patient’s X-ray anatomical positions was 11.38 mm. The registered model exhibited a high degree of alignment with the curvature of the spine in X-ray i

关键词： Three dimensional computer graphics

来源：评论

学校读者我要写书评

暂无评论

GVI: Guideable Visual Interpretation on medical tomographic images to improve the performance of deep network

引用

Pattern Recognition Letters 2025年 196卷 162-168页

作者： Hui Liu Fan Wei Lixin Yan Sushan Wang Chongfu Jia Lina Zhang Jiansheng Peng Yi Xu School of Control Science and Engineering Dalian University of Technology Dalian China Key Laboratory of Intelligent Control and Optimization for Industrial Equipment of Ministry of Education Dalian University of Technology Dalian China Liaoning Key Laboratory of Integrated Circuit and Biomedical Electronic System Dalian China Department of Cardiovascular Radiology The First Affiliated Hospital of Dalian Medical University Dalian China Department of Radiology The First Affiliated Hospital of Dalian Medical University Dalian China Department of Artificial Intelligence and Manufacturing Hechi University Hechi China

In medical image analysis, the demand for interpretable deep neural networks is rapidly growing. However, a major challenge is that most existing interpretative methods are applied after training, leading to a lack of integration with the model’s learning process. As a result, these methods often fail to highlight regions within complex medical images critical for decision-making, such as abnormal tissues or lesions, which are essential for accurate diagnoses and treatment planning. This paper introduces Guided Visual Interpretation (GVI), a framework designed to enhance both the performance and interpretability of deep networks. Building on a deep network model with image-level labels, GVI incorporates a small amount of pixel-level annotations combined with attention mechanisms. These mechanisms facilitate visual interpretation through forward propagation, directing the model’s focus to the most relevant regions. By aligning the network’s decision-making with human cognitive processes, GVI improves interpretability. In our study, an attention layer was added after the convolutional layers of a pre-trained classification network. GVI is trained using a mixed supervision approach that integrates pixel-level annotations with a large amount of image-level data. Experimental results on both private and public datasets show that GVI generates visual explanations consistent with human decision-making principles and achieves superior classification accuracy compared to traditional methods. These findings highlight GVI’s potential to improve interpretability and diagnostic performance in critical fields like medical imaging.

关键词： Class activation maps Image classification Interpretability Mixed supervision

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：