检索结果-内蒙古大学图书馆

arXiv 2019年

作者： Liu, Zihan Huang, Bo Cui, Yuqi Xu, Yifan Zhang, Bo Zhu, Lixia Wang, Yang Jin, Lei Wu, Dongrui Key Laboratory of Ministry of Education for Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China Reproductive Medicine Center Huazhong University of Science and Technology Tongji Hospital Wuhan430074 China

Time-lapse is a technology used to record the development of embryos during in-vitro fertilization (IVF). Accurate classification of embryo early development stages can provide embryologists valuable information for assessing the embryo quality, and hence is critical to the success of IVF. This paper proposes a multi-task deep learning with dynamic programming (MTDL-DP) approach for this purpose. It first uses MTDL to pre-classify each frame in the time-lapse video to an embryo development stage, and then DP to optimize the stage sequence so that the stage number is monotonically non-decreasing, which usually holds in practice. Different MTDL frameworks, e.g., oneto-many, many-to-one, and many-to-many, are investigated. It is shown that the one-to-many MTDL framework achieved the best compromise between performance and computational cost. To our knowledge, this is the first study that applies MTDL to embryo early development stage classification from time-lapse videos. Copyright © 2019, The Authors. All rights reserved.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

On the periodic dynamics of memristor-based neural networks with leakage and time-varying delays

引用

Neurocomputing 2017年 219卷 163-173页

作者： Ping Jiang Zhigang Zeng Jiejie Chen School of Automation Huazhong University of Science and Technology Wuhan 430074 China School of Computer Hubei PolyTechnic University Huangshi 435002 China Key Laboratory of Image Processing and Intelligent Control of Education Ministry of China Wuhan 430074 China School of Computer Science and Technology Hubei Normal University Huangshi 435000 China

In this paper, the periodic dynamics have been studied for a general kind of memristor-based neural networks with leakage and time-varying delays. Some new sufficient conditions have been derived ensuring that the existence, uniqueness and globally exponential stability of the periodic solution for the neural network by using differential inclusions theory, the topological degree theory in set-valued analysis and Lyapunov function technique and so on. As a special case, we have shown that the existence, uniqueness and global exponential stability of equilibrium point for the autonomous neural networks with leakage delays.

关键词： Memristor-based neural networks Leakage delays Time-varying delays Periodic solution Global exponential stability

来源：评论

学校读者我要写书评

暂无评论

Zero-Sum Game Based Secure Tracking control of UAV Against FDI Attacks Using Fixed-Time Convergent Reinforcement Learning

Zero-Sum Game Based Secure Tracking Control of UAV Against F...

引用

IEEE Conference on Decision and control

作者： Zhenyu Gong Feisheng Yang Dongrui Wu Research & Development Institute of Northwestern Polytechnical University in Shenzhen School of Automation Northwestern Polytechnical University Xi'an China Key Laboratory of the Ministry of Education for Image Processing and Intelligent Control School of Artificial and Automation Huazhong University of Science and Technology Wuhan China

In this paper, a fixed-time convergent reinforcement learning (RL) algorithm is developed to realize the secure tracking control of the unmanned aerial vehicle (UAV) via the zero-sum game for the first time. To mitigate FDI attack on actuators that may cause the UAV to deviate from the reference trajectory, a zero-sum differential game framework is built in which the secure controller tries to minimize the common performance function, yet the attacker plays a contrary role. Obtaining the optimal secure tracking controller depends on solving the Hamilton-Jacobi-Isaacs (HJI) equation related to the zero-sum game. Therefore, a critic-only online RL algorithm is proposed that can converge in a fixed time interval, with the corresponding convergence proof provided. A simulation example is given to show the effectiveness of the raised method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Pcda: Efficient Decision-Based Black-Box Attack for Point Clouds

SSRN

引用

SSRN 2023年

作者： Liu, Lianguang Lv, Junting Li, Depeng Xu, Bingrong Chen, Chongyang Zeng, Zhigang School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China Key Laboratory of Image Processing and Intelligent Control of Education Ministry of China Wuhan430074 China School of Automatio Wuhan University of Technology Wuhan430074 China

Deep neural networks for 3D point clouds have shown significant progress in safety-critical applications. However, the robustness of these deep 3D models is insufficient-explored when faced with adversarial attacks. The existing point clouds attack methods usually rely on detailed model information (white-box attacks) or access to the training data (transfer-based black-box attacks). In this work, we make the first attempt in decision-based black-box attack for point clouds (PCDA), which only leverages the predicted labels of the victim models. Specifically, a novel perturbation projection based on chamfer distance is developed to be compatible with 3D irregular data, and the gradient estimation based on low-dimensional subspaces can generate smoother perturbations. Moreover, armed with iterative closest point, the novel perturbation initialization can improve the effectiveness of targeted black-box attacks. Evaluated on a variety of typical 3D models, our algorithm can efficiently generate smooth 3D adversarial samples with a lower magnitude of perturbations. © 2023, The Authors. All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Visual SLAM Based on Dynamic Object Detection

Visual SLAM Based on Dynamic Object Detection

引用

第33届中国控制与决策会议

作者： Bocheng Chen Gang Peng Dingxin He Cheng Zhou Bin Hu School of Artificial Intelligence and Automation Huazhong University of Science and Technology Key Laboratory of Image Processing and Intelligent Control Ministry of Education School of Civil and Hydraulic Engineering Huazhong University of Science and Technology Shantui Construction Machinery Co. Ltd.

On the one hand,traditional visual SLAM does not consider dynamic objects in the scene,on the other hand,deep learning technology has been widely used in computer *** paper combines the two organically,and proposes an algorithm that uses dynamic object detection to improve the robustness of visual SLAM in a dynamic ***,we use the object detection network integrated into the attention mechanism to detect the dynamic target in the key ***,we follow the optical flow detection to further determine the dynamic feature points in the scene and eliminate ***,we use the static feature points for camera tracking to achieve highly robust monocular visual *** method described in this paper can not only eliminate dynamic feature points,but also retain as many static feature points as *** method described in this paper is compared with the original ORB-SLAM2 algorithm and DS-SLAM algorithm,and tested with public data *** results show that the method described in this paper can effectively eliminate the influence of dynamic objects on the visual SLAM algorithm.

关键词： Visual SLAM dynamic object detection deep learning optical flow detection attention mechanism

来源：评论

学校读者我要写书评

暂无评论

UniAnimate: Taming Unified Video Diffusion Models for Consistent Human image Animation

arXiv

引用

arXiv 2024年

作者： Wang, Xiang Zhang, Shiwei Gao, Changxin Wang, Jiayu Zhou, Xiaoqiang Zhang, Yingya Yan, Luxin Sang, Nong Key Laboratory of Ministry of Education for Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China Alibaba Group Hangzhou310052 China University of Science and Technology of China Hefei230026 China

Recent diffusion-based human image animation techniques have demonstrated impressive success in synthesizing videos that faithfully follow a given reference identity and a sequence of desired movement poses. Despite this, there are still two limitations: i) an extra reference model is required to align the identity image with the main video branch, which significantly increases the optimization burden and model parameters;ii) the generated video is usually short in time (e.g., 24 frames), hampering practical applications. To address these shortcomings, we present a UniAnimate framework to enable efficient and long-term human video generation. First, to reduce the optimization difficulty and ensure temporal coherence, we map the reference image along with the posture guidance and noise video into a common feature space by incorporating a unified video diffusion model. Second, we propose a unified noise input that supports random noised input as well as first frame conditioned input, which enhances the ability to generate long-term video. Finally, to further efficiently handle long sequences, we explore an alternative temporal modeling architecture based on state space model to replace the original computation-consuming temporal Transformer. Extensive experimental results indicate that UniAnimate achieves superior synthesis results over existing state-of-the-art counterparts in both quantitative and qualitative evaluations. Notably, UniAnimate can even generate highly consistent one-minute videos by iteratively employing the first frame conditioning strategy. Code and models will be publicly available. Project page: https://***/. Copyright © 2024, The Authors. All rights reserved.

关键词： Animation

来源：评论

学校读者我要写书评

暂无评论

ROAD CENTRAL CONTOUR EXTRACTION FROM HIGH RESOLUTION SATELLITE image USING TENSOR VOTING FRAMEWORK

ROAD CENTRAL CONTOUR EXTRACTION FROM HIGH RESOLUTION SATELLI...

引用

2006 International Conference on Machine Learning and Cybernetics(IEEE第五届机器学习与控制论坛)

作者： SHENG ZHENG JIAN LIU WEN-ZHONG SHI GUANG-XI ZHU Electronic & Information Engineering Dept. Huazhong University of Science and Technology Wuhan 430 State Education Commission Key Laboratory for Image Processing and Intelligent Control Inst.for Pat Advanced Research Center for Spatial Information Technology Department of Land Surveying and Geo-In Chinese Key lab of optic & electronics Huazhong University of Science and Technology Wuhan 430074

In this paper, a unique road contour extraction approach from high resolution satellite image is proposed, in which the road contour was extracted in two steps. Firstly, support vector machines (SVM) was employed merely to classify the image into two groups of categories: a road group and a non-road group. The identified road group images are the discrete and irregularly distributed sampled points, and they are an uncompleted data set for the road. Secondly, the road contour was extracted from the road group images using the tensor voting framework, since the tensor voting technique is superior to the traditional methods in extracting the geometrical structure from the uncompleted data set. The experimental results on the high resolution satellite image demonstrate that the proposed approach worked well with images comprised by both rural and urban area features.

关键词： Tensor voting framework Road central line extraction High-resolution satellite image

来源：评论

学校读者我要写书评

暂无评论

A nonlinear sliding mode observer for the estimation of temperature distribution in the planar solid oxide fuel cell stack

A nonlinear sliding mode observer for the estimation of temp...

引用

第五届世界氢能技术大会(WHTC2013)

作者： Huan Cheng Xi Li Jianhua Jiang Lin Zhang Jian Li Jie Yang Department of control science and Engineering Key Laboratory of Education Ministry for Image Processing and Intelligent Control Huazhong University of Science and Technology Wuhan 430074 China School of Materials Science and Engineering State Key Laboratory of Material Processing and Die and Mould Technology Huazhong University of Science and Technology Wuhan 430074 China School of Mechanical and Electronic Information China University of Geosciences Wuhan 430074 China

来源：评论

学校读者我要写书评

暂无评论

Adaptive flocking of multi-agent systems with locally Lipschitz nonlinearity

Adaptive flocking of multi-agent systems with locally Lipsch...

引用

Chinese control Conference (CCC)

作者： Ningzi Zhang Housheng Su Michael Z. Q. Chen Xiaofan Wang Department of Control Science and Engineering Image Processing and Intelligent Control Key Laboratory of Education Ministry of China Huazhong University of Science and Technology Wuhan China Department of Mechanical Engineering University of Hong Kong Hong Kong China Department of Automation Shanghai JiaoTong University Shanghai China

ISBN: (纸本)9781467325813

This paper investigates adaptive flocking of multi-agent systems (MASs) with a virtual leader. All agents and the virtual leader share the same intrinsic nonlinear dynamics, which satisfies a locally Lipschitz condition and depends on both position and velocity information of the agent itself. Under the assumption that the initial network is connected, an approach to preserving the connectivity of the network is proposed. Based on the Lyapunov stability theory, an adaptive flocking control law is derived to make the MASs track the virtual leader without collision. Finally, a numerical example is presented to illustrate the effectiveness of the theoretical results.

关键词： Multiagent systems Artificial neural networks Nonlinear dynamical systems educational institutions Heuristic algorithms Couplings

来源：评论

学校读者我要写书评

暂无评论

A multi-label front propagation approach for object segmentation

A multi-label front propagation approach for object segmenta...

引用

International Conference on Pattern Recognition

作者： H. Li A. Elmoataz J. Fadili S. Ruan Department of Electronics & Information Engineering State Education Commission Laboratory for Image Processing and Intelligent Control Huazhong University of Science and Technology China GREYC-ISMRA CNRS Caen France LUSAC Cherbourg-Octeville France Equipe Image L.A.M.(EA 2075) Department GE&II IUT de Troyes Troyes France

For effective image segmentation methods, speed, accuracy and smoothness of the result are essential. In this paper, an iterative object segmentation approach is proposed based on minimal path theory. Each iterative step includes one morphological dilatation and one multi-label front propagation. A narrow band is obtained by dilating the current contour with the known size. A new contour is again formed by multi-label front propagation, which is based on minimal path theory. Its propagation speed is decided by the local image mean values together with the edge function. The final boundary is obtained automatically through finite iterations. This algorithm is a global optimization method. It is simple and fast with complexity O(N). The initial contour may be chosen freely. The multi-label front propagation guarantees continuity and smooth contours with the capability to handle topology changes. Furthermore, it is easy to extend to the 3D case. Some experimental results are also presented.

关键词： Object segmentation image segmentation Active contours Topology Computer vision Robust stability control engineering education Laboratories image processing intelligent control

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：