检索结果-南通市图书馆

Training and Testing Object Detectors With Virtual Images

在线全文

学校读者我要写书评

暂无评论

IEEE/CAA Journal of Automatica Sinica 2018年第2期5卷 539-546页

作者： Yonglin Tian Xuan Li kunfeng wang Fei-Yue wang Department of Automation University of Science and Technology of China Hefei 230027 China State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijirtg 100190 China School of Automation Beijing Institute of Technology Beijing 100081. China Qingdao Academy of Intelligent Industries Qingdao 266000 China Research Center for Computational Experiments and Parallel Systems Technology National University of Defense Technology Changsha 410073 China IEEE

In the area of computer vision, deep learning has produced a variety of state-of-the-art models that rely on massive labeled data. However, collecting and annotating images from the real world is too demanding in terms of labor and money investments, and is usually inflexible to build datasets with specific characteristics, such as small area of objects and high occlusion level. Under the framework of Parallel Vision, this paper presents a purposeful way to design artificial scenes and automatically generate virtual images with precise annotations.A virtual dataset named Parallel Eye is built, which can be used for several computer vision tasks. Then, by training the DPM(Deformable parts model) and Faster R-CNN detectors, we prove that the performance of models can be significantly improved by combining Parallel Eye with publicly available real-world datasets during the training phase. In addition, we investigate the potential of testing the trained models from a specific aspect using intentionally designed virtual datasets, in order to discover the flaws of trained models. From the experimental results, we conclude that our virtual dataset is viable to train and test the object detectors.

关键词： Index Terms--Deep learning object detection parallel vision virtual dataset

Prediction of methane storage in covalent organic frameworks using big-data-mining approach

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Chemical Engineering 2021年第11期34卷 286-296页

作者： Huan Zhang Peisong Yang Duli Yu kunfeng wang Qingyuan Yang State Key Laboratory of Organic-Inorganic Composites Beijing University of Chemical TechnologyBeijing 100029China Beijing Advanced Innovation Center for Soft Matter Science and Engineering Beijing University of Chemical TechnologyBeijing 100029China College of Information Science and Technology Beijing University of Chemical TechnologyBeijing 100029China

A combination of computational materials screening and machine learning(ML)technique is being adopted as a popular approach to study various materials toward application of *** this work,we began with high-throughput molecular simulations to calculate the methane storage(6.5 MPa)and deliverable(6.5-0.58 MPa)capacities of 404,460 covalent organic frameworks(COFs)at 298 ***,the full data sets with 23 features were randomly split into training and test sets in a ratio of 20:80,which were applied to evaluate the prediction abilities of several ML algorithms,including gradient boosting decision tree(GBDT),neural network(NN),support vector machine(SVM),random forest(RF)and decision tree(DT).The results indicate that the RF model has the highest prediction accuracy,which was further employed to reduce the dimension of features space and quantitatively analyze the relative importance of each feature *** binary classification predictors built using the features with the highest influence weight can give a successful identification of top-performing candidates from the test set containing 323,168 COFs with an accuracy exceeding 96%.The deliverable capacities of the identified COFs were found to outperform those reported so far for various *** findings may provide a useful guidance for the design and synthesis of new high-performance materials for methane storage application.

关键词： Covalent organic framework Monte Carlo simulation Methane Machine learning Model

Generative Adversarial Networks:Introduction and Outlook

在线全文

学校读者我要写书评

暂无评论

IEEE/CAA Journal of Automatica Sinica 2017年第4期4卷 588-598页

作者： kunfeng wang Chao Gou Yanjie Duan Yilun Lin Xinhu Zheng Fei-Yue wang IEEE State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Qingdao Academy of Intelligent Industries State Key Laboratory of Management and Control for Complex Systems Institute of AutomationChinese Academy of Sciences University of Chinese Academy of Sciences Department of Computer Science and Engineering University of Minnesota Research Center for Computational Experiments and Parallel Systems Technology National University of Defense Technology

Recently, generative adversarial networks(GANs)have become a research focus of artificial intelligence. Inspired by two-player zero-sum game, GANs comprise a generator and a discriminator, both trained under the adversarial learning *** goal of GANs is to estimate the potential distribution of real data samples and generate new samples from that *** their initiation, GANs have been widely studied due to their enormous prospect for applications, including image and vision computing, speech and language processing, etc. In this review paper, we summarize the state of the art of GANs and look into the future. Firstly, we survey GANs' proposal background,theoretic and implementation models, and application ***, we discuss GANs' advantages and disadvantages, and their development trends. In particular, we investigate the relation between GANs and parallel intelligence,with the conclusion that GANs have a great potential in parallel systems research in terms of virtual-real interaction and integration. Clearly, GANs can provide substantial algorithmic support for parallel intelligence.

关键词： ACP approach adversarial learning generative adversarial networks(GANs) generative models parallel intelligence zero-sum game

A decouple-decomposition noise analysis model for closed-loop mode-localized tilt sensors

在线全文

学校读者我要写书评

暂无评论

Microsystems & Nanoengineering 2023年第6期9卷 297-307页

作者： kunfeng wang XingYin Xiong Zheng wang Liangbo Ma BoWen wang WuHao Yang Xiaorui Bie ZhiTian Li XuDong Zou The State Key Laboratory of Transducer Technology Aerospace Information Research InstituteChinese Academy of SciencesBeijingChina School of Electronic Electrical and Communication EngineeringUniversity of Chinese Academy of SciencesBeijingChina Ming Hsieh Department of Electrical and Computer Engineering University of Southern CaliforniaLos AngelesCAUSA Shangdong Key Laboratory of Low-altitude Airspace Surveillance Network Technology QiLu Aerospace Information Research InstituteJinanChina

The development of mode-localized sensors based on amplitude output metrics has attracted increasing attention in recent years due to the potential of such sensors for high sensitivity and ***-localization phenomena leverage the interaction between multiple coupled resonant modes to achieve enhanced performance,providing a promising solution to overcome the limitations of traditional sensing *** noise plays a key role in determining the resolution of mode-localized sensors,as the output metric is derived from the measured AR(amplitude ratio)within the weakly coupled resonator ***,the amplitude noise originating from the weakly coupled resonator's closed-loop circuit has not yet been fully *** paper presents a decouple-decomposition(DD)noise analysis model,which is applied to achieve high resolution in a mode-localized tilt sensor based on a weakly coupled resonator closed-loop *** DD noise model separates the weakly coupled resonators using the decoupling method considering the nonlinearity of the *** integrating the decoupled weakly coupled resonators,the model decomposes the weakly coupled resonator's closed-loop circuit into distinct paths for amplitude and phase noise *** DD noise model reveals noise effects at various circuit nodes and models the system noise in the closed-loop circuit of the weakly coupled ***/Simulink simulations verify the model's accuracy when compared to theoretical *** the optimal operating point,the mode-localized tilt sensor achieves an input-referred instability of 3.91×10^(-4°)and an input-referred AR of PSD of 2.01×10^(-4°)⁄√Hz using the closed-loop noise *** model is also applicable to other varieties of mode-localized sensors.

关键词： weakly couple overcome

基于偶极子天线辐射反向创建空间球形聚焦光斑

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

光学学报 2024年第10期44卷 428-435页

作者：王俊杰杨艳芳何英李琦王昆峰上海大学理学院物理系上海200444

与传统的仅能生成光轴方向的球形聚焦光斑不同,本文提出了一种在4Pi聚焦系统中通过反聚焦偶极子天线辐射场产生具有规定空间方向和间距的球形聚焦光斑的方法。该方法是将规定好长度和极化方向的空间偶极子天线置于4Pi聚集系统的焦点处,... 详细信息

与传统的仅能生成光轴方向的球形聚焦光斑不同,本文提出了一种在4Pi聚焦系统中通过反聚焦偶极子天线辐射场产生具有规定空间方向和间距的球形聚焦光斑的方法。该方法是将规定好长度和极化方向的空间偶极子天线置于4Pi聚集系统的焦点处,通过逆问题解析求解出生成球形聚焦光斑的物镜瞳孔面上的输入场。然后使用瞳孔面处的场,并选择合适的偶极子天线长度,就能获得球形聚焦光斑。数值结果表明,创建的球形聚焦光斑的空间方向与设置的偶极子天线的极化方向一致,球形聚焦光斑之间的距离也等于偶极子天线的长度。本文提出的方法比传统方法更灵活,可以创建具有规定空间位置的球形聚焦光斑,这对空间任意位置捕获纳米粒子具有很大的应用价值。

关键词：物理光学球形聚焦光斑偶极子天线 4Pi聚焦系统

在线全文

学校读者我要写书评

暂无评论

自适应视觉惯性地磁紧耦合定位系统

光学精密工程 2024年第7期32卷 1023-1033页

作者：付鹏万振华王琨锋赵开春清华大学精密仪器系北京100084 北京信息科技大学仪器科学与光电工程学院北京100192 广西大学机械工程学院广西南宁530004

为了解决视觉惯性组合导航系统绝对航向信息缺失与姿态发散的问题,进一步提升系统定位精度,设计并实现了针对未知磁场环境下的自适应视觉惯性地磁紧耦合定位系统。基于常用的三轴磁力计内外参数联合标定原理,提出了一种地磁信息全局及... 详细信息

为了解决视觉惯性组合导航系统绝对航向信息缺失与姿态发散的问题,进一步提升系统定位精度,设计并实现了针对未知磁场环境下的自适应视觉惯性地磁紧耦合定位系统。基于常用的三轴磁力计内外参数联合标定原理,提出了一种地磁信息全局及帧间约束残差构建方法。通过磁强度差异动态调整融合权重,利用非线性优化方法设计视觉/惯性/地磁紧耦合系统实现自身运动状态估计。最后,在校园环境下开展户外实验。实验结果表明:该系统能够在部分建筑及车辆磁场干扰环境下稳定运行,定位精度均优于0.8%(RMSE),相比于VINS位置误差平均降低了约24%,且具有良好的实时性。磁力计及自适应融合方法的引入可以有效改善现有视觉惯性导航系统的定位性能,为无人系统提供高可靠性的实时定位结果。

关键词：组合导航地磁场自适应系统非线性优化

维普期刊数据库博看期刊

在线全文

学校读者我要写书评

暂无评论

一种高灵敏度低噪声硅微谐振式加速度计

中国惯性技术学报 2021年第3期29卷 281-287页

作者：邹旭东熊兴崟汪政王坤锋杨伍昊李志天传感技术国家重点实验室中国科学院空天信息创新研究院北京100190 齐鲁空天信息研究院济南250132

MEMS加速度计经过近四十年的发展,是目前产业化最为成功、应用最为广泛的MEMS器件之一。以硅微机械谐振器作为敏感元件的谐振式MEMS加速度计因具有检测精度高、线性度好、量程大、抗环境噪声能力强等优点,成为新一代高性能MEMS加速度计... 详细信息

MEMS加速度计经过近四十年的发展,是目前产业化最为成功、应用最为广泛的MEMS器件之一。以硅微机械谐振器作为敏感元件的谐振式MEMS加速度计因具有检测精度高、线性度好、量程大、抗环境噪声能力强等优点,成为新一代高性能MEMS加速度计的重要发展方向。针对微小型无人平台的长时惯性导航、姿态测量等需求,设计了一种具有增敏结构的硅微谐振式加速度计,通过改进微杠杆转轴与惯性质量块支撑梁的几何形状并利用有限元仿真方法进行参数优化,在不增加芯片面积的前提下有效提升了器件灵敏度。器件设计量程±50 g,采用集成圆片级真空封装的SOI-MEMS工艺制造并配套设计了基于0.35 mm工艺的接口ASIC电路用于实现加速度计的闭环工作。所研制的原理样机测试表明,加速度计敏感谐振器品质因数为29300,灵敏度630.81 Hz/g,噪声≤1.7μg/√Hz,零偏不稳定性(Allan方差)≤2.3μg。

关键词： MEMS加速度计谐振式加速度计圆片级真空封装低噪声加速度计

在线全文

学校读者我要写书评

暂无评论

基于卷积神经网络的三维目标检测研究综述

模式识别与人工智能 2021年第12期34卷 1103-1119页

作者：王亚东田永林李国强王坤峰李大字北京化工大学信息科学与技术学院北京100029 中国科学技术大学自动化系合肥230022 中国科学院自动化研究所复杂系统管理与控制国家重点实验室北京100190

深度学习尤其卷积神经网络为精确目标检测提供可能,推动三维目标检测在自动驾驶、机器人等领域发挥重要作用.文中综述基于卷积神经网络的三维目标检测研究进展.首先总结三维目标检测的应用价值、基本流程及存在的挑战.再介绍卷积神经网... 详细信息

深度学习尤其卷积神经网络为精确目标检测提供可能,推动三维目标检测在自动驾驶、机器人等领域发挥重要作用.文中综述基于卷积神经网络的三维目标检测研究进展.首先总结三维目标检测的应用价值、基本流程及存在的挑战.再介绍卷积神经网络基本原理、典型的二维目标检测网络结构、常用的开源数据集及点云表示形式等相关基础知识.然后介绍卷积神经网络在三维目标检测中的应用进展,根据不同数据模态及方法共性对方法进行梳理.最后对当前三维目标检测研究存在的问题进行论述,对未来的研究发展趋势进行展望.

关键词：卷积神经网络三维目标检测点云多模态融合

在线全文

学校读者我要写书评

暂无评论

平行图像:图像生成的一个新型理论框架

模式识别与人工智能 2017年第7期30卷 577-587页

作者：王坤峰鲁越王雨桐熊子威王飞跃中国科学院自动化研究所复杂系统管理与控制国家重点实验室北京100190 青岛智能产业技术研究院平行视觉创新技术中心青岛266000 中国科学院大学计算机与控制学院北京100049 国防科学技术大学军事计算实验与平行系统技术研究中心长沙410073

为了提高计算机视觉系统的泛化能力,要求利用大规模、多样化、带标注的图像数据集,对视觉模型进行充分的学习与评估.由于从实际场景中获取图像具有局限性,文中提出一种图像生成理论框架,称为平行图像.平行图像的核心单元是软件定义的人... 详细信息

为了提高计算机视觉系统的泛化能力,要求利用大规模、多样化、带标注的图像数据集,对视觉模型进行充分的学习与评估.由于从实际场景中获取图像具有局限性,文中提出一种图像生成理论框架,称为平行图像.平行图像的核心单元是软件定义的人工图像系统.从实际场景中获取特定的图像"小数据",输入人工图像系统,生成大量新的人工图像数据.文中总结平行图像的实现方法,包括图形渲染、图像风格迁移、生成式模型等,并且对比分析人工图像和实际图像的特点,讨论领域适应策略.

关键词：平行图像模型学习图形渲染图像风格迁移生成式模型