检索结果-内蒙古大学图书馆

No-Reference Image Quality Assessment using Meta-SGD

学校读者我要写书评

暂无评论

No-Reference Image Quality Assessment using Meta-SGD

Chinese Automation Congress (CAC)

作者： Qingqing Yan Bei Dong Xin Li Yiwen Xiang Dapeng Luo Longsheng Wei School of Automation China University of Geosciences Wuhan China Key Laboratory of Geological Survey and Evaluation of Ministry of Education Wuhan China Hubei key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan China Intelligent Technology co. Ltd of Chinese Construction Third Engineering Bureau Wuhan China School of Electronic Information and Mechanics China University of Geosciences Wuhan China

ISBN: (纸本)9781665426480

Deep learning-based no-reference image quality assessment (NR-IQA) algorithm requires massive training data and labels, but data scarcity usually exists in practical applications. Most of the previous metrics used pre-trained networks to solve this problem by fine-tuning on the target tasks. Unfortunately, the network trained on other tasks cannot be directly applied to IQA tasks and failed when encountering complex distortions. To reduce dependence on labeled data and improve generalization ability of unknown distortion types, a NR-IQA metric based on Meta-SGD is proposed, which can learn the quick adaptability of humans and obtain general meta-knowledge for distortion evaluation. Specifically, we collected a great many NR-IQA tasks with different distortion types to pre-train the model; then a meta-learning metric based on optimization, Meta-SGD, is proposed to acquire the meta-knowledge when evaluating images quality with various known distortions; finally, the obtained pre-trained meta-model can be directly applied to the new NR-IQA tasks only by fine-tuning a few images. Experiments on the TID2013 database demonstrate that the model after meta-learning obviously outperforms than some existing methods, and the average SRCC value for all types of distortion performance tests can reach 0.91.

关键词： Measurement Image quality Adaptation models Automation Databases Training data Distortion

Deep Rank-Consistent Pyramid Model for Enhanced Crowd Counting

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Gao, Jiaqi Huang, Zhizhong Lei, Yiming Shan, Hongming Wang, James Z. Wang, Fei-Yue Zhang, Junping The Shanghai Key Laboratory of Intelligent Information Processing School of Computer Science Fudan University Shanghai200433 China The Institute of Science and Technology for Brain-inspired Intelligence MOE Frontiers Center for Brain Science Fudan University Shanghai200433 China The Shanghai Center for Brain Science and Brain-Inspired Technology Shanghai201210 China The College of Information Sciences and Technology The Pennsylvania State University University ParkPA16802 United States The State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China The Institute of Systems Engineering Macau University of Science and Technology China Qingdao Academy of Intelligent Industries Qingdao266109 China

Most conventional crowd counting methods utilize a fully-supervised learning framework to establish a mapping between scene images and crowd density maps. They usually rely on a large quantity of costly and time-intensive pixel-level annotations for training supervision. One way to mitigate the intensive labeling effort and improve counting accuracy is to leverage large amounts of unlabeled images. This is attributed to the inherent self-structural information and rank consistency within a single image, offering additional qualitative relation supervision during training. Contrary to earlier methods that utilized the rank relations at the original image level, we explore such rank-consistency relation within the latent feature spaces. This approach enables the incorporation of numerous pyramid partial orders, strengthening the model representation capability. A notable advantage is that it can also increase the utilization ratio of unlabeled samples. Specifically, we propose a Deep Rank-consistEnt pyrAmid Model (DREAM), which makes full use of rank consistency across coarse-to-fine pyramid features in latent spaces for enhanced crowd counting with massive unlabeled images. In addition, we have collected a new unlabeled crowd counting dataset, FUDAN-UCC, comprising 4,000 images for training purposes. Extensive experiments on four benchmark datasets, namely UCF-QNRF, ShanghaiTech PartA and PartB, and UCF-CC-50, show the effectiveness of our method compared with previous semi-supervised methods. The codes are available at https://***/bridgeqiqi/DREAM. Copyright © 2022, The Authors. All rights reserved.

关键词： Image enhancement

Overlapped Human Pose Estimation using Non-Maximum Suppression based on Shape Similarity

学校读者我要写书评

暂无评论

Overlapped Human Pose Estimation using Non-Maximum Suppressi...

Chinese Automation Congress (CAC)

作者： Xuefu Yu Jinsheng Li Fei Xiao Wei Liu Dapeng Luo Longsheng Wei School of Automation China University of Geosciences Wuhan China Key Laboratory of Geological Survey and Evaluation of Ministry of Education Wuhan China Hubei key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan China Intelligent Technology co. Ltd of Chinese Construction Third Engineering Bureau Wuhan China School of Electronic Information and Mechanics China University of Geosciences Wuhan China

ISBN: (纸本)9781665426480

Deep Convolutional Neural Networks (DNNs) have made significant progress on human pose estimation task in recent years. However, the question of whether we are making the most use of the information we get from DNNs has not been well studied. Top-down approach which first detects person body box then utilizes single person pose estimation by heatmap is a widely used method in multi-person pose estimation. However, information that predicted by DNNs contains more information than we consider. In person box prediction process, when two persons are closed to each other, the Non-Maximum Suppression (NMS) we often use filters out the body boxes with smaller scores. For tackling this problem, we apply NMS based on Shape Similarity to person body boxes detection for tackling the problem that deleting the overlapped body boxes with smaller scores, and then extract keypoint coordinates and tag information for filter noise candidates from heatmap by Trasformer to improve the prediction accuracy. At last we develop an algorithm for refining keypoint coordinate by tag information. In experiment, we got mean Average Precision(mAP) of 76.0% in COCO2017 validation dataset, better than 75.6% with TransPose.

关键词： Heating systems Shape Pose estimation Refining Filtering algorithms Information filters Prediction algorithms

An Universal control Strategy for Planar 2-DoF Underactuated Manipulator with One Passive Joint

学校读者我要写书评

暂无评论

An Universal Control Strategy for Planar 2-DoF Underactuated...

第40届中国控制会议

作者： Zixin Huang Yaosheng Zhou Zhen Chen Wei Wang Lejun Wang School of Electrical and Information Engineering Wuhan Institute of Technology Hubei Key Laboratory of Digital Textile Equipment Wuhan Textile University Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education School of Innovation and Entrepreneurship Wuhan Institute of Technology School of Information and Safety Engineering Zhongnan University of Economics and Law

This paper presents an unified control strategy based on the trajectory planning and tracking control for the planar 2-Do F underactuated manipulator. The trajectory of the active link is composed of two parts. The first part trajectories are designed according to the initial and target angles of the active link. The second part trajectories are designed based on the constraints between the underactuated link and the active link. Meanwhile, the parameters of the trajectories are optimized by the differential evolution algorithm(DEA) to ensure all links eventually reach to their target values by tracking the designed trajectories. Then,the sliding mode variable structure controller are designed to make the active link track their trajectories. The effectiveness of such strategy is demonstrated through simulation results.

关键词： Planar 2-DoF underactuated manipulator Trajectory planning Tracking control Differential evolution algorithm

A Bioinspired retinal neural network for accurately extracting small-target motion information in cluttered backgrounds

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Huang, Xiao Qiao, Hong Li, Hui Jiang, Zhihong School of Mechatronical Engineering Advanced Innovation Center for Intelligent Robots and Systems Key Laboratory of Biomimetic Robots and Systems of Chinese Ministry of Education Beijing Institute of Technology Beijing100081 China State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China Beijing Key Laboratory of Research and Application for Robotic Intelligence of’Hand-Eye-Brain’ Interaction Beijing100190 China University of Chinese Academy of Sciences Beijing100049 China

—Robust and accurate detection of small moving targets in cluttered moving backgrounds is a significant and challenging problem for robotic visual systems to perform search and tracking tasks. Inspired by the neural circuitry of elementary motion vision in the mammalian retina, this paper proposes a bioinspired retinal neural network based on a new neurodynamics-based temporal filtering and multiform 2-D spatial Gabor filtering. This model can estimate motion direction accurately via only two perpendicular spatiotemporal filtering signals, and respond to small targets of different sizes and velocities through adjusting the dendrite field size of spatial filter. Meanwhile, an algorithm of directionally selective inhibition is proposed to suppress the target-like features in the moving background, which can reduce the influence of background motion effectively. Extensive synthetic and real-data experiments show that the proposed model works stably for small targets of a wider size and velocity range, and has better detection performance than other bioinspired models. Additionally, it can also extract the information of motion direction and motion energy accurately and rapidly. Copyright © 2021, The Authors. All rights reserved.

关键词： Ophthalmology

EDOLAB: An Open-Source Platform for Education and Experimentation with Evolutionary Dynamic optimization Algorithms

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Peng, Mai Yazdani, Delaram She, Zeneng Yazdani, Danial Luo, Wenjian Li, Changhe Branke, Juergen Nguyen, Trung Thanh Gandomi, Amir H. Yang, Shengxiang Jin, Yaochu Yao, Xin School of Automation China University of Geosciences Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Engineering Research Center of Intelligent Technology for Geo-Exploration Ministry of Education Wuhan China Research Institute Faculty of Engineering and Technology Liverpool John Moores University United Kingdom School of Computer Science and Technology Harbin Institute of Technology China Guangdong Provincial Key Laboratory of Novel Security Intelligence Technologies School of Computer Science and Technology Harbin Institute of Technology Peng Cheng Laboratory China School of Artificial Intelligence Anhui University of Sciences & Technology China Information Systems Management and Analytics in Warwick Business School University of Warwick United Kingdom Faculty of Engineering & Information Technology University of Technology Sydney Australia Obuda University Hungary School of Computer Science and Informatics De Montfort University United Kingdom Department of Artificial Intelligence School of Engineering Westlake University China School of Computer Science University of Birmingham United Kingdom

Many real-world optimization problems exhibit dynamic characteristics, posing significant challenges for traditional optimization techniques. Evolutionary Dynamic optimization Algorithms (EDOAs) are designed to address these challenges effectively. However, in existing literature, the reported results for a given EDOA can vary significantly. This inconsistency often arises because the source codes for many EDOAs, which are typically complex, have not been made publicly available, leading to error-prone re-implementations. To support researchers in conducting experiments and comparing their algorithms with various EDOAs, we have developed an open-source MATLAB platform called the Evolutionary Dynamic optimization laboratory (EDOLAB). This platform not only facilitates research but also includes an educational module designed for instructional purposes. The education module allows users to observe: a) a 2-dimensional problem space and its morphological changes following each environmental change, b) the behaviors of individuals over time, and c) how the EDOA responds to environmental changes and tracks the moving optimum. The current version of EDOLAB features 25 EDOAs and four fully parametric benchmark generators. © 2023, CC BY-NC-SA.

关键词： optimization algorithms

A Bioinspired approach-sensitive neural network for collision detection in cluttered and dynamic backgrounds

学校读者我要写书评

暂无评论

arXiv 2021年

—Rapid, accurate and robust detection of looming objects in cluttered moving backgrounds is a significant and challenging problem for robotic visual systems to perform collision detection and avoidance tasks. Inspired by the neural circuit of elementary motion vision in the mammalian retina, this paper proposes a bioinspired approach-sensitive neural network (ASNN) that contains three main contributions. Firstly, a direction-selective visual processing module is built based on the spatiotemporal energy framework, which can estimate motion direction accurately via only two mutually perpendicular spatiotemporal filtering channels. Secondly, a novel approach-sensitive neural network is modeled as a push-pull structure formed by ON and OFF pathways, which responds strongly to approaching motion while insensitivity to lateral motion. Finally, a method of directionally selective inhibition is introduced, which is able to suppress the translational backgrounds effectively. Extensive synthetic and real robotic experiments show that the proposed model is able to not only detect collision accurately and robustly in cluttered and dynamic backgrounds but also extract more collision information like position and direction, for guiding rapid decision making. © 2021, CC BY.

关键词： Mammals

H∞ Consensus control for Discrete-Time Stochastic Multi-agent systems with Infinite Markov Jumps

学校读者我要写书评

暂无评论

H∞ Consensus Control for Discrete-Time Stochastic Multi-age...

Annual Conference of Industrial Electronics Society

作者： Jie Wang Xin Chen School of Automation China University of Geosciences Wuhan P. R. China Hubei Key Laboratory of Advanced Control and Intelligent Automation for Complex Systems

ISBN: (数字)9781728154145

ISBN: (纸本)9781728154152

This paper focuses on multi-agent systems under the disturbances of parameter uncertainties, noises and external disturbances simultaneously, which are more suitable for the changing environments than most ideal models. For this case, the robust H∞ consensus control problem is studied to attenuate the influence of disturbances below a given level. We aim to design a consensus controller such that the closed-loop muti-agent systems reach the desired H ∞ performance. By tools of Kronecker product, graph theory, and infinite horizon bounded real lemma, sufficient conditions for existence of H ∞ consensus protocol design are obtained in terms of linear matrix inequalities (LMIs). Finally, a numerical example is given to show the validity of the method.

关键词： Multi-agent systems Symmetric matrices Markov processes Stochastic systems Robot sensing systems Eigenvalues and eigenfunctions Topology

Multi-UAV 3D Path Planning in Simultaneous Attack

学校读者我要写书评

暂无评论

Multi-UAV 3D Path Planning in Simultaneous Attack

International Conference on control and Automation (ICCA)

作者： Chuyi Xiong Bin Xin Miao Guo Yulong Ding Hao Zhang School of Automation Beijing Institute of Technology Beijing China State Key Laboratory of Intelligent Control and Decision of Complex Systems Beijing Institute of Technology Beijing China

ISBN: (数字)9781728190938

ISBN: (纸本)9781728190945

As the key technology of UAV mission planning, path planning has a very important impact on the combat effectiveness of UAV. In order to destroy enemy units, multiple UAVs are often needed to attack the target at the same time. In the complex combat environment, the path planning of multiple UAVs is a very complex problem. On the basis of genetic algorithm, this paper designs a path planning algorithm to ensure that multiple UAVs arrive at the same time from different angles on the premise of satisfying environment constraints and the motion constraints of UAV. The algorithm uses the method of regionalization to initialize the path, designs a reasonable fitness function, and adds an adaptive disturbance operator to plan the path of each UAV. Finally, the length difference between the paths planned for each UAV is taken as the path evaluation factor. Some of the paths are replanned to achieve the goal of the same path length and simultaneous attack by UAV. The effectiveness of the algorithm is verified by experiments in typical combat scenarios, and the superiority of the adaptive disturbance operator in the optimization of the algorithm is verified by the contrast experiment.

关键词： Path planning Radar Planning optimization Radar detection Genetic algorithms Three-dimensional displays