检索结果-内蒙古大学图书馆

Slim-RFFNet: Slim deep convolution random Fourier feature network for image classification

KNOWLEDGE-BASED SYSTEMS 2022年 237卷 107878-107878页

作者： Wang, Tingting Dong, Bo Zhang, Kaichun Li, Junbao Xu, Lei Harbin Inst Technol Fac Comp Harbin 150001 Heilongjiang Peoples R China Harbin Inst Technol Dept Control Sci & Engn Harbin 150001 Heilongjiang Peoples R China Second Acad CASIC X Lab Beijing 100854 Peoples R China Osaka Electrocommun Univ Neyagawa Osaka 5728530 Japan Harbin Inst Technol Sch Elect & Informat Engn Harbin 150001 Heilongjiang Peoples R China

The standard kernel method is computationally expensive because it needs to store and compute the inverse of the Gram matrix. Furthermore, the classification accuracy of the single kernel method is not effective or efficient as the method's ability to extract features. The random Fourier feature method establishes a connection between the kernel method and deep learning to solve the above problems. In this paper, we propose a novel slim deep random Fourier feature network named Slim-RFFNet, which introduces convolution into kernel learning. We use the hierarchical strategy and skip connection to construct a deep network structure and lighten the model by using quantization. Experiments conducted on classification benchmarks MNIST and CIFAR10 demonstrate that the proposed SlimRFFNet significantly outperforms current state-of-the-art deep kernel learning methods. Our algorithm also achieves a trade-off between accuracy and latency. The proposed network can be applied to resource-constrained embedded AI devices. The experimental results on the edge computing system show that our algorithm has a small memory footprint and fast inference speed on small edge devices, and thus meets the requirements for practical applications. (c) 2021 Elsevier B.V. All rights reserved.

关键词： Deep kernel learning Random Fourier features Convolution Quantization Classification

来源：评论

学校读者我要写书评

暂无评论

QUBIQ: Uncertainty Quantification for Biomedical image Segmentation Challenge

arXiv

引用

arXiv 2024年

作者： Li, Hongwei Bran Navarro, Fernando Ezhov, ivan Bayat, Amirhossein Das, Dhritiman Kofler, Florian Shit, Suprosanna Waldmannstetter, Diana Paetzold, Johannes C. Hu, Xiaobin Wiestler, Benedikt Zimmer, Lucas Amiranashvili, Tamaz Prabhakar, Chinmay Berger, Christoph Weidner, Jonas Alonso-Basanta, Michelle Rashid, Arif Baid, Ujjwal Adel, Wesam Alis, Deniz Baheti, Bhakti Bai, Yingbin Bhat, Ishaan Cetindag, Sabri Can Chen, Wenting Cheng, Li Dutande, Prasad Dular, Lara Elattar, Mustafa A. Feng, Ming Gao, Shengbo Huisman, Henkjan Hu, Weifeng Innani, Shubham Ji, Wei Karimi, Davood Kuijf, Hugo J. Kwak, Jin Tae Le, Hoang Long Li, Xiang Lin, Huiyan Liu, Tongliang Ma, Jun Ma, Kai Ma, Ting Oksuz, Ilkay Holland, Robbie Oliveira, Arlindo L. Pal, Jimut Bahan Pei, Xuan Qiao, Maoying Saha, Anindo Selvan, Raghavendra Shen, Linlin Silva, Joao Lourenco Spiclin, Ziga Talbar, Sanjay Wang, Dadong Wang, Wei Wang, Xiong Wang, Yin Xi, Ruiling Xu, Kele Yang, Yanwu Yergin, Mert Yu, Shuang Zeng, Lingxi Zhang, YingLin Zhao, Jiachen Zheng, Yefeng Zukovec, Martin Do, Richard Becker, Anton Simpson, Amber Konukoglu, Ender Jakab, Andras Bakas, Spyridon Joskowicz, Leo Menze, Bjoern Department of Informatics Technical University of Munich Germany Athinoula A. Martinos Center for Biomedical Imaging Massachusetts General Hospital Harvard Medical School United States Department of Quantitative Biomedicine University of Zurich Switzerland University Children’s Hospital Zurich University of Zurich Switzerland Department of Radioncology and Radiation Theraphy Klinikum rechts der Isar Technical University of Munich Germany Department of Information Technology and Electrical Engineering ETH-Zurich Switzerland Department of Radiology Memorial Sloan Kettering Cancer Center New York City United States Department of Biomedical and Molecular Sciences Queen’s University Canada TranslaTUM - Central Institute for Translational Cancer Research Technical University of Munich Germany McGovern Institute Massachusetts Institute of Technology United States Institute for Diagnostic and Interventional Radiology Unveristy Zurich Hospital Switzerland BioMedIA Imperial College London United Kingdom Department of Radiation Oncology University of Pennsylvania PA United States University of Pennsylvania PA United States Department of Radiation Oncology Winship Cancer Institute of Emory University Georgia United States Nile University Cairo Egypt Department of Medical Sciences Acibadem University Istanbul Turkey Shri Guru Gobind Singhji Institute of Engineering and Technology Maharashtra Nanded India Trustworthy Machine Learning Lab University of Sydney Australia Image Sciences Institute University Medical Center Utrecht Netherlands Computer Engineering Department Istanbul Technical University Istanbul Turkey School of Computer Science Shenzhen University Shenzhen China University of Alberta United States University of Ljubljana Faculty of Electrical Engineering Ljubljana Slovenia Tongji University Shanghai China OPPO Research Institute Shanghai China School of Biological and Medical Engineering Beihang University Beijing China Harvard Medical School Boston

Uncertainty in medical image segmentation tasks, especially inter-rater variability, arising from differences in interpretations and annotations by various experts, presents a significant challenge in achieving consistent and reliable image segmentation. This variability not only reflects the inherent complexity and subjective nature of medical image interpretation but also directly impacts the development and evaluation of automated segmentation algorithms. Accurately modeling and quantifying this variability is essential for enhancing the robustness and clinical applicability of these algorithms. We report the set-up and summarize the benchmark results of the Quantification of Uncertainties in Biomedical image Quantification Challenge (QUBIQ), which was organized in conjunction with International Conferences on Medical image Computing and Computer-Assisted Intervention (MICCAI) 2020 and 2021. The challenge focuses on the uncertainty quantification of medical image segmentation which considers the omnipresence of inter-rater variability in imaging datasets. The large collection of images with multi-rater annotations features various modalities such as MRI and CT;various organs such as the brain, prostate, kidney, and pancreas;and different image dimensions 2D-vs-3D. A total of 24 teams submitted different solutions to the problem, combining various baseline models, Bayesian neural networks, and ensemble model techniques. The obtained results indicate the importance of the ensemble models, as well as the need for further research to develop efficient 3D methods for uncertainty quantification methods in 3D segmentation tasks. © 2024, CC BY-NC-ND.

关键词： image segmentation

来源：评论

学校读者我要写书评

暂无评论

Resampling method for reliability-based design optimization based on thermodynamic integration and parallel tempering

引用

MECHANICAL SYSTEMS AND SIGNAL processing 2021年 156卷 107630-107630页

作者： Cheng, Kai Lu, Zhenzhou Xiao, Sinan Zhang, Xiaobo Oladyshkin, Sergey Nowak, Wolfgang Northwestern Polytech Univ Sch Aeronaut Xian 710072 Peoples R China Univ Stuttgart Dept Stochast Simulat & Safety Res Hydrosyst IWS Pfaffenwaldring 5a D-70569 Stuttgart Germany Forschungszentrum Julich Inst Bio & Geosci IBG Agrosphere 3 Julich Germany

In this paper, a fully decoupled simulation method is proposed for reliability-based design optimization (RBDO) based on thermodynamic integration and parallel tempering (TIPT). We show that the failure probability function and its gradient can be obtained simultane-ously with once generalized reliability analysis, and thus the RBDO problem is converted to the traditional optimization problem efficiently. Firstly, the design parameters are deemed as uniformly distributed random variables, and an auxiliary probability density function (PDF) of random design variables is constructed to cover its whole parameter space. Then, based on thermodynamic integration, the estimation of failure probability is con-verted to a series of simple integration problems with smooth integrand, and they are esti-mated by running multiple Markov chains using the so-called parallel tempering method. Finally, importance sampling (IS) is used to estimate the failure probability function and its gradient, and the IS samples are obtained by resampling from the existing Markov chains without extra computation. The proposed method is tested with severa benchmarks, and the results show that it provides robust solution for problems with various nonlinear con-straints compared to other popular methods, include double-loop Monte Carlo simulation (MCS), Quantile MCS, sequential optimization and reliability assessment, performance measure approach and reliability index approach. (c) 2021 Elsevier Ltd. All rights reserved.

关键词： Reliability-based design optimization Resampling Thermodynamic integration parallel tempering

来源：评论

学校读者我要写书评

暂无评论

Object Pose Estimation for Robotic Grasping based on Multi-view Keypoint Detection

Object Pose Estimation for Robotic Grasping based on Multi-v...

引用

IEEE International Conference on Big Data and Cloud Computing (BdCloud)

作者： Zheyuan Hu Renluan Hou Jianwei Niu Xiaolong Yu Tao Ren Qingfeng Li State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and Engineering Beihang University Beijing China Hangzhou Innovation Institute Beihang University Hangzhou China Zhengzhou University Research Institute of Industrial Technology School of Information Engineering Zhengzhou University Zhengzhou China

Industrial robots can replace human labour to perform a variety of tasks. Among these tasks, robotic grasping is the most primary industrial robot operation. However, conventional robotic grasping methods could become inapplicable for cluttered and occluded objects. To address the issue, we adopt object pose estimation (OPE) to facilitate robotic grasping of cluttered and occluded objects and propose an object detection model based on 2D-RGB multi-view features. The proposed model is built by adding four transpose convolution layers into the Resnet backbone to obtain desirable 2D feature maps of object keypoints in each image. In addition, we design a feature-fusion model to produce 3D coordinates of keypoints from 2D multi-view features based on the volumetric aggregation method, along with a keypoint-detection confidence of each view to assist the optimality judgment of the robotic grasping. Extensive experiments are conducted to verify the accuracy of OPE, and the experimental results indicate the substantial performance improvements of the proposed approach over conventional methods in various scenarios.

关键词： Solid modeling Three-dimensional displays Service robots image color analysis Robot kinematics Pose estimation Robot vision systems

来源：评论

学校读者我要写书评

暂无评论

Standard Deviation Based Adaptive Gradient Compression For distributed Deep Learning 20

Standard Deviation Based Adaptive Gradient Compression For D...

引用

20th IEEE/ACM International Symposium on Cluster, Cloud and Internet Computing (CCGrid)

作者： Chen, Mengqiang Yan, Zijie Ren, Jiangtao Wu, Weigang Sun Yat Sen Univ Sch Data & Comp Sci Guangzhou Peoples R China

ISBN: (纸本)9781728160955

distributed deep learning has been proposed to train large scale neural network models with huge amounts of datasets using multiple workers. Since workers need to frequently communicate with each other to exchange gradients for updating parameters, communication overhead is always a major challenge in distributed deep learning. To cope with this challenge, gradient compression has been used to reduce the amount of data to be exchanged. However, existing compression methods, including both gradient quantization and gradient sparsification, either hurt model performance significantly or suffer from inefficient compression. In this paper, we propose a novel approach, called Standard Deviation based Adaptive Gradient Compression (SDAGC), which can simultaneously achieve model training with low communication overhead and high model performance in synchronous training. SDAGC uses the standard deviation of gradients in each layer of the neural network to dynamically calculate a suitable threshold according to the training process. Moreover, several associated methods, including residual gradient accumulation, local gradient clipping, adaptive learning rate revision and momentum compensation, are integrated to guarantee the convergence of the model. We verify the performance of SDAGC on various machine learning tasks: image classification, language modeling and speech recognition. The experiment results show that, compared with other existing works, SDAGC can achieve a gradient compression ratio from 433x to 2021x with similar or even better accuracy.

关键词： distributed deep learning gradient compression standard deviation neural networks parallel processing

来源：评论

学校读者我要写书评

暂无评论

一种基于Spark的公路裂缝图像处理方法

引用

计算机与数字工程 2022年第8期50卷 1795-1801页

作者：李杨刘庆华刘东华黄凯枫贾叙文江苏科技大学镇江212000

随着公路修建里程增长,为了更高效率地解决路面养护问题,对路面裂缝检测技术进行研究。论文将传统的图像分割处理与Spark结合,提出了一种基于Spark的路面裂缝图像分割处理方法。先将图像预处理转化为二进制文件,读入到分布式文件系统;... 详细信息

随着公路修建里程增长,为了更高效率地解决路面养护问题,对路面裂缝检测技术进行研究。论文将传统的图像分割处理与Spark结合,提出了一种基于Spark的路面裂缝图像分割处理方法。先将图像预处理转化为二进制文件,读入到分布式文件系统;然后通过传递函数,对裂缝进行图像分割来获得裂缝区域;最后,以局部自适应阈值算法结合Spark平台来进行图像的并行处理,完成裂缝的高效率提取。以在校园拍摄的路面图像数据为实验数据,分别在处理时间和检测准确率这两方面进行实验。由结果可知:利用Spark对数千张图像数据进行计算时,在确定千余张图像数据规模时,双节点比单节点处理时间更少,最多减少了20.4%。在图像分割的准确率方面,已经由之前方法的47.5%提高到了86.3%。实验结果表明,基于Spark的裂缝图像处理,结合局部自适应阈值分割算法进行数千张裂缝图像分割,能有效提升传统方法的处理效率,能实现大规模图像的并行处理。

关键词：公路裂缝图像处理 Spark 并行处理大数据局部自适应阈值分割

来源：评论

学校读者我要写书评

暂无评论

A Raster Data Framework Based on distributed Heterogeneous Cluster

引用

JOURNAL OF THE INDIAN SOCIETY OF REMOTE SENSING 2019年第4期47卷 715-723页

作者： Bhushan, R. Phani Somayajulu, D. V. L. N. Venkatraman, S. Subramanyam, R. B. V. ADRIN Dept Space Hyderabad Telangana India NIT Dept Comp Sci & Engn Warangal Andhra Pradesh India

Advancements in satellite imaging and sensor technologies result in capturing of large amount of spatial data. Many parallel processing techniques based on data or control parallelism have been attempted during the past 2 decades to provide performance improvement in image processing applications such as urban sprawl, weather prediction and crop estimation. These techniques have used block-based distributed file processing or the more modern MapReduce-based programming for implementation which still have gaps between optimal and best processing in terms of resource scheduling, data distribution and ease of programming. In this paper, we present a layered framework for parallel data processing to improve storage, retrieval and processing performance of spatial data on an underlying distributed file system. The paper presents a data placement strategy across a distributed HDFS cluster in a way to optimize spatial data retrieval and processing. The presence of neighborhood pixels local to the processing node in a distributed environment reduces network latencies and improves the efficiency of applications such as object recognition, change detection and site selection. We evaluate the data placement strategy on a four-node HDFS cluster and show that it can deliver good performance benefits by way of reading blocks of data at almost 10-12 times the default, which contributes to the improvement in efficiency of the various applications that use region growing methods.

关键词： Big data Remote sensing Spatial MapReduce HDFS RecordReader Raster data Vector data Resource scheduler Retrieval Engine

来源：评论

学校读者我要写书评

暂无评论

Study of large axial space rebinning algorithm in γ-photon industrial tomography image reconstruction

引用

JOURNAL OF INSTRUMENTATION 2021年第12期16卷

作者： Wang, M. Zhao, M. Yao, M. Liu, J. Guo, R. Nanjing Univ Aeronaut & Astronaut Coll Automat Engn 29 Gen Rd Nanjing 211106 Peoples R China

The accuracy of the existing single slice and Fourier rebinning algorithms depends on the projection angle of the line of response. The increase of such projection angle with the detector size, typical in the large axial space of gamma-photon industrial detection, and the loss of some projection data after rebinning, result in the degradation of the image quality. In addition, those algorithms consider the probability of positron annihilation equally distributed along the line of response, which prevents to estimate accurately the positions of the annihilation point, and can originate artifacts and noise in the reconstructed image. In this work, we propose an alternative large axial space rebinning algorithm. In that algorithm, initially the line of response is divided into transverse and axial components. Then, each line of response is uniformly rebinned into all the 2D sinogram data intersecting with it. To improve the accuracy of the estimate of the annihilation point location and suppress the noise effectively, we assign a Gaussian weight coefficient to the projection data, and optimise the rebinning algorithm with it. Finally, we reconstruct the image on the basis of the 2D sinograms with the optimised weights. On the computational side, the algorithm is also accelerated by making use of parallel computing. Both simulation and experimental results show that the proposed method improves the contrast and spatial resolution of 2D reconstructed images. Furthermore, the reconstruction time is not affected by the new method, which is therefore expected to meet the demand of gamma-photon industrial inspection imaging.

关键词： Data processing methods Detection of defects Inspection with gamma rays

来源：评论

学校读者我要写书评

暂无评论

Evaluation of suturing performance in general surgery and ocular microsurgery by combining computer vision-based software and distributed fiber optic strain sensors: a proof-of-concept

引用

INTERNATIONAL JOURNAL OF COMPUTER ASSISTED RADIOLOGY AND SURGERY 2020年第8期15卷 1359-1367页

作者： Handelman, Amir Keshet, Yariv Livny, Eitan Barkan, Refael Nahum, Yoav Tepper, Ronnie Holon Inst Technol Fac Engn Dept Elect Engn Holon Israel Clalit Hlth Serv Simultech Ctr Med Simulat Meir Hosp Med Ctr Kefar Sava Israel Rabin Med Ctr Dept Ophthalmol Petah Tiqwa Israel Tel Aviv Univ Sackler Fac Med Tel Aviv Israel

Purpose Improper suturing may cause an inadequate wound healing process and wound dehiscence as well as infection and even graft rejection in case of corneal transplantation. Hence, training surgeons in correct suturing procedures and objectively assessing their surgical skills is desirable. methods Two complementary methods for assessment of suturing skills in two medical fields (general surgery and ocular microsurgery) were demonstrated. Suturing quality is assessed by computer vision software. Evaluation of stitching flow of operation is based on measuring strain induced in an optical fiber that is placed in proximity to the wound and parallel thereto and is pressed and passed by wound stitches. Results Our software generated a score for suturing outcome in both general surgery and ocular microsurgery when the stitching was done on a patch. Every trainee received a score in the range 0-100 that describes his/her performance. Strain values were recognized when using a patch in general surgery and a rubber patch in ocular microsurgery, but were less distinct in (disqualified) human cornea. Conclusions We proved a concept of an objective scoring method (based on various image processing algorithms) for assessment of suturing performance. It was also shown that fiber optic strain sensors are sensitive to the flow of stitching operation on a patch but are less sensitive to the flow of stitching operation on a human cornea. By combining these two methods, we can comprehensively evaluate the suturing performance objectively.

关键词： Sutures Fiber optic sensors Computer vision Ocular microsurgery Surgical education

来源：评论

学校读者我要写书评

暂无评论

Underwater image Enhancement based on Dehazing and Color Correction 17

Underwater Image Enhancement based on Dehazing and Color Cor...

引用

IEEE Int Conf on parallel and distributed processing with Applications, Big Data and Cloud Computing, Sustainable Computing and Communications, Social Computing and Networking (ISPA/BDCloud/SocialCom/SustainCom)

作者： Li, Hanyu Zhuang, Peixian Wang, Wei Li, Jingjing Nanjing Univ Informat Sci & Technol Nanjing Peoples R China

ISBN: (纸本)9781728143286

Clear and clean underwater images can provide valuable information, which are crucial for developing, exploring, and protecting the underwater resources. However, the raw underwater image seldom fulfills the requirements concerning the underwater research due to the serious image degradation. To enhance single underwater image, we propose an effective and new method combined on the dehazing and color correction algorithm. First, we obtain the dehazed image by a fusion dehazing method calculated on the difference between maximum green-blue dark channels and maximum red dark channel, and the top 0.1% bright pixels in green-blue dark channels. Second, we obtain the enhanced image via the color restoration method corresponding to the human visual system. Finally, we use an efficient and simple weight fusion strategy to incorporate the dehazed image and the enhanced image for yielding the final high quality underwater image. Subjective assessment demonstrates that our method can remove haze, correct color cast, improve image brightness, and preserve image naturalness. The enhanced results obtain the highest values in terms of entropy, average gradient, UCIQE, UIQM, and CCF, which is superior to several existing underwater enhancement methods. Moreover, our method can improve other degraded image quality, such as low-light image and haze image, and can benefit the application tests, such as key points matching and edge detection.

关键词： Single underwater image enhancement image dehazing retinex color correction

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：