检索结果-内蒙古大学图书馆

Towards to robust and generalized medical image segmentation framework

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Chen, Yurong Zhang, Hui Wang, Yaonan Liu, Lizhu Wu, Q.M. Jonathan Yang, Yimin National Engineering Laboratory of Robot Visual Perception and Control Technology School of Robotics Hunan University Hunan Changsha410082 China Department of Electrical and Computer Engineering University of Windsor WindsorONN9B3P4 Canada Department of Computer Science Lakehead University Vector Institute TorontoM5G 1M1 Canada

—To mitigate the radiologist’s workload, computer-aided diagnosis with the capability to review and analyze medical images is gradually deployed. Deep learning-based region of interest segmentation is among the most exciting use cases. However, this paradigm is restricted in real-world clinical applications due to the poor robustness and generalization. The issue is more sinister with a lack of training data. In this paper, we address the challenge from the representation learning point of view. We investigate that the collapsed representations, as one of the main reasons which caused poor robustness and generalization, could be avoided through transfer learning. Therefore, we propose a novel two-stage framework for robust generalized segmentation. In particular, an unsupervised Tile-wise AutoEncoder (T-AE) pretraining architecture is coined to learn meaningful representation. Furthermore, this learned prior knowledge is transferred to the segmentation benchmark. Coupled with an image reconstruction network, the representation keeps to be decoded, encouraging the model to capture more semantic features. Experiments of lung segmentation on multi chest X-ray datasets are conducted. Empirically, the comprehensive results demonstrate the superior generalization capability of the proposed framework on unseen domains in terms of high performance and robustness to corruption, especially under the scenario of the limited training data. © 2021, CC BY.

关键词： Computer aided diagnosis

Machine learning–based robust trajectory tracking control for FSGR

学校读者我要写书评

暂无评论

Machine learning–based robust trajectory tracking control f...

7th International Symposium on Test Automation and Instrumentation, ISTAI 2018

作者： Jia, Lin Wang, Yaonan Zhang, Changfan Zhao, Kaihui Zhou, Langming College of Electrical and Information Engineering Hunan University Changsha China National Engineering Laboratory for Robot Visual Perception and Control Hunan University Changsha China College of Electrical and Information Engineering Hunan University of technology Zhuzhou China

Here, a robust adaptive trajectory tracking algorithm is proposed for free-form surface grinding robot (FSGR) in metal surface production line. Machine-learning method is used for robot dynamic approximation which is hard to obtain directly. Adaptive law is proposed to adjust the neural network parameters. Sliding-mode control is employed to deal with the disturbance, joint friction, and approximation error of the adaptive machine learning. The scheme based on machine-learning feedforward compensation can significantly reduce the chattering of sliding mode. The performance of the proposed control scheme is illustrated through simulations. © 2019 Institution of engineering and technology. All rights reserved.

关键词： Machine learning

Dynamic Update Siamese Networks with Deeper Features

学校读者我要写书评

暂无评论

Dynamic Update Siamese Networks with Deeper Features

作者： Jianhao Tan Shanshan Tan Wang Yin Liming Liu Siyuan Zhang Yaonan Wang College of Electrical and Information Engineering Hunan University National Engineering Laboratory for Robot Visual Perception and Control Technology

Visual object tracking has been a concern topic these years,and many trackers have achieved good results in various *** researches and breakthroughs have made many improvements to solve problems such as drift,lighting,deformation and *** this paper,we improve the structure of the AlexNet[1] network by designing the three important influencing factors of the receptive field size,total network step size,and feature filling of the twin *** from this,we add a smoothing matrices and a background suppression matrices to effectively learn the features of the first few frames as much as *** multilayer feature elements can learn online about target appearance changes and background suppression,and we train them by using continuous video sequences.

关键词： Target tracking Siamese network Residual network Dynamic update

A fast and robust visual detection method with binary template matching for glass bottle body

学校读者我要写书评

暂无评论

A fast and robust visual detection method with binary templa...

第三十八届中国控制会议

作者： Xianen Zhou Yaonan Wang Qing Zhu The College of Electrical and Information Engineering Hunan University The National Engineering Laboratory for Robot Visual Perception and Control Technology

As multiple glass bottle bodies appear in the same one image from a view, the high precision and speed localization of the bottle body is difficult for the traditional visual detection methods. To overcome the problem, the imaging system is first briefly presented. And a new visual detection framework, named BTMfast, for glass bottle body with binary template matching is proposed. First, the input image is down-sampled to reduce the computational complexity. The subimages containing body region are obtained according to the position priors. Then, the neck or body of a bottle is taken as the template, and the template image and each sumimage is segmented with OTSU segmentation algorithm. Thirdly, the central line of the body in the original image can be obtained by the binary template matching. Finally, the top boundary of the bottle mouth is enhanced by a filtering operation with a new kernel. And the mouth position is obtained by segmenting the filtered image and column scanning. Two datasets are created for evaluating the performance of the proposed method. Compared with many conventional approaches on the two datasets, our method can achieve the optimal result. The localization error is about 1.5, the consumed time is about 4.3 ms.

关键词： Visual detection template matching localization glass bottle body

Target Tracking Based on Correlation Filter for Scale Offset

学校读者我要写书评

暂无评论

Target Tracking Based on Correlation Filter for Scale Offset

作者： Jianhao Tan Xiaoping Ma Yaonan Wang College of Electrical and Information Engineering Hunan University National Engineering Laboratory for Robot Visual Perception and Control Technology

Existing scale estimation methods mainly use fixed-size tracking box for target ***,when the moving direction of the target changes from near to far or from far to near,the size of the current tracking box cannot adapt to the scale *** order to overcome the disadvantage of invariant size of tracking box in the scale change,this paper proposes a scale offset estimation method based on correlation filter by combining the state of previous frame and current frame of the *** the scale proposal box in the scale layer,and adjust the position and size of the actual tracking box in real time according to the size of the proposal box,target position and scale *** this paper,OTB-100 is used as the *** obtained tracking results show that our algorithm Ours has better tracking performance than the existing others tracking algorithm in scale variation,and it has better tracking effect and stronger robustness in drift and occlusion events.

关键词： frame Target Tracking Based on Correlation Filter for Scale Offset Filter

Adaptive finite-time synchronized control of multi-robotic fiber placement system with model uncertainties and disturbances

学校读者我要写书评

暂无评论

ISA Transactions 2025年

作者： Ronghua Zhang Yaonan Wang Wenfang Xie Pengcheng Li Haoran Tan Yiming Jiang College of Electrical and Information Engineering Hunan University Changsha 410082 Hunan China National Engineering Laboratory of Robot Visual Perception and Control Technology Hunan University Changsha 410082 Hunan China Department of Mechanical Industrial and Aerospace Concordia University Montreal H3G2W1 Quebec Canada College of Mechanical and Electrical Engineering Nanjing University of Aeronautics and Astronautics Nanjing 210016 Jiangsu China School of Robotics Hunan University Changsha 410082 Hunan China

The use of multiple robots to manufacture composite components represents a critical development direction for fiber placement systems (FPSs). In multi-robotic fiber placement systems (MRFPSs) with heterogeneous mechanical structures, robots collaborate to perform fiber placement tasks. Consequently, robot synchronization emerges as a primary factor in determining the performance of the fiber placement process. However, the difficulty in establishing accurate system models and the presence of disturbances are two significant challenges to achieving precise robot synchronization. Additionally, the system is expected to exhibit desirable dynamic characteristics, such as finite-time error convergence. To address these issues and requirements, we propose a novel adaptive finite-time synchronization control (AFSC) algorithm for the system. Specifically, a finite-time sliding mode observer is developed to handle kinematic uncertainty. A novel fast non-singular terminal sliding mode (FNTSM) manifold is constructed in the AFSC algorithm. Moreover, the control algorithm integrates an adaptive law to handle dynamic uncertainty and an adaptive term to counteract disturbances. Performance analysis demonstrates that the AFSC ensures that the coupled, synchronization, and tracking errors converge to zero within finite time. Furthermore, simulations and experiments are conducted to validate the effectiveness of the AFSC algorithm.

关键词： Disturbances Fiber placement Finite-time control Model uncertainties Multi-robots Synchronization control

Projected Sliced Wasserstein Autoencoder-based Hyperspectral Images Anomaly Detection

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Chen, Yurong Zhang, Hui Wang, Yaonan Wu, Q. M. Jonathan Yang, Yimin National Engineering Laboratory of Robot Visual Perception and Control Technology School of Robotics Hunan University Hunan Changsha410082 China Department of Electrical and Computer Engineering University of Windsor WindsorONN9B3P4 Canada Department of Computer Science Lakehead University ONP7B 5E1 Canada Vector Institute TorontoONM5G 1M1 Canada

Anomaly detection refers to identifying the observation that deviates from the normal pattern, which has been an active research area in various domains. Recently, the increasing data scale, complexity, and dimension turns the traditional representation and statistical-based outlier detection method into challenging. In this paper, we leverage the generative model in hyperspectral images anomaly detection. The gist is to model the distribution of the normal data, while the out-of-distribution sample can be viewed as the outlier. At first, the variational inference-based anomaly detection methods are investigated. We theoretically and empirically find that they are unstable due to the strong notion of distance (f-divergence) served as the regularization. Secondly, this paper introduces sliced Wasserstein distance, which is a weaker distribution measure compared with f-divergence. However, the number of randomly slicing poses a difficulty to estimate the true distance. In the end, we propose a projected sliced Wasserstein (PSW) autoencoder-based anomaly screening method. In particular, we leverage a computation-friendly eigen-decomposition method to find the principal component as slicing the high-dimensional data. Furthermore, our proposed distance can be calculated with the closed-form, even the prior distribution is not Gaussian. Comprehensive experiments conducted on various real-world hyperspectral anomaly detection benchmarks demonstrate the superior performance of our proposed method. © 2021, CC BY.

关键词： Spectroscopy

P-WAE: Generalized patch-wasserstein autoencoder for anomaly screening

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Zhang, Hui Chen, Yurong Wang, Yaonan Cao, Yihong Wu, Q.M. Jonathan Yang, Yimin National Engineering Laboratory of Robot Visual Perception and Control Technology School of Robotics Hunan University Changsha Hunan410082 China Department of Electrical and Computer Engineering University of Windsor WindsorONN9B3P4 Canada Department of Computer Science Lakehead University OntarioP7B 5E1 Canada Vector Institute TorontoM5G 1M1 Canada

To mitigate the inspector’s workload and improve the quality of the product, computer vision-based anomaly detection (AD) techniques are gradually deployed in real-world industrial scenarios. Recent anomaly analysis benchmarks progress to generative models. The aim is to model the defect-free distribution so that anomalies can be classified as out-of-distribution samples. Nevertheless, there are two disturbing factors that need researchers and deployers to prioritize: (i) the simplistic prior latent distribution inducing limited expressive capability;(ii) the collapsed mutual-dependent features resulting in poor generalization. In this paper, we propose a novel Patch-wise Wasserstein AutoEncoder (P-WAE) architecture to alleviate those challenges. In particular, a patch-wise variational inference model coupled with solving the jigsaw puzzle is designed, which is a simple yet effective way to increase the expressiveness and complexity of the latent manifold. This alleviates the blurry reconstruction problem. In addition, the Hilbert-Schmidt Independence Criterion (HSIC) bottleneck is introduced to constrain the over-regularization representation. Comprehensive experiments, conducted on the MVTec AD dataset, demonstrate the superior performance of our proposed method. © 2021, CC BY.

关键词： Anomaly detection

MyoPS: A Benchmark of Myocardial Pathology Segmentation Combining Three-Sequence Cardiac Magnetic Resonance Images

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Li, Lei Wu, Fuping Wang, Sihan Luo, Xinzhe Martín-Isla, Carlos Zhai, Shuwei Zhang, Jianpeng Liu, Yanfei Zhang, Zhen Ankenbrand, Markus J. Jiang, Haochuan Zhang, Xiaoran Wang, Linhong Arega, Tewodros Weldebirhan Altunok, Elif Zhao, Zhou Li, Feiyan Ma, Jun Yang, Xiaoping Puybareau, Elodie Oksuz, Ilkay Bricq, Stephanie Li, Weisheng Punithakumar, Kumaradevan Tsaftaris, Sotirios A. Schreiber, Laura M. Yang, Mingjing Liu, Guocai Xia, Yong Wang, Guotai Escalera, Sergio Zhuang, Xiahai School of Data Science Fudan University Shanghai China School of Biomedical Engineering Shanghai Jiao Tong University Shanghai China Departament de Matemàtiques & Informàtica Universitat de Barcelona Barcelona Spain Computer Vision Center Universitat Autònoma de Barcelona Spain School of Mechanical and Electrical Engineering University of Electronic Science and Technology of China Chengdu China School of Computer Science and Engineering Northwestern Polytechnical University Xi'An China College of Electrical and Information Engineering Hunan University Changsha China National Engineering Laboratory for Robot Visual Perception and Control Technology Changsha China College of Physics and Information Engineering Fuzhou University Fuzhou China Comprehensive Heart Failure Center Wuerzburg University Hospitals Wuerzburg Germany School of Engineering University of Edinburgh Edinburgh United Kingdom School of Robotics Xi'an Jiaotong-Liverpool University Suzhou China Department of Electrical and Computer Engineering University of California Los Angeles United States Department of Radiology and Diagnostic Imaging University of Alberta Edmonton Canada Chongqing Key Laboratory of Image Cognition Chongqing University of Posts and Telecomm-Unications Chongqing China ImViA Laboratory Université Bourgogne Franche-Comté Dijon France Computer Engineering Department Istanbul Technical University Istanbul Turkey Le Kremlin-Bicêtre France Department of Mathematics Nanjing University of Science and Technology Nanjing China Department of Mathematics Nanjing University Nanjing China

Assessment of myocardial viability is essential in diagnosis and treatment management of patients suffering from myocardial infarction, and classification of pathology on myocardium is the key to this assessment. This work defines a new task of medical image analysis, i.e., to perform myocardial pathology segmentation (MyoPS) combining three-sequence cardiac magnetic resonance (CMR) images, which was first proposed in the MyoPS challenge, in conjunction with MICCAI 2020. The challenge provided 45 paired and pre-aligned CMR images, allowing algorithms to combine the complementary information from the three CMR sequences for pathology segmentation. In this article, we provide details of the challenge, survey the works from fifteen participants and interpret their methods according to five aspects, i.e., preprocessing, data augmentation, learning strategy, model architecture and post-processing. In addition, we analyze the results with respect to different factors, in order to examine the key obstacles and explore potential of solutions, as well as to provide a benchmark for future research. We conclude that while promising results have been reported, the research is still in the early stage, and more in-depth exploration is needed before a successful application to the clinics. Note that MyoPS data and evaluation tool continue to be publicly available upon registration via its homepage (***/zhuangxiahai/0/myops20/). © 2022, CC BY-NC-SA.

关键词： Magnetic resonance imaging