检索结果-内蒙古大学图书馆

Modeling Inter-Intra Heterogeneity for Graph Federated Learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Yu, Wentao Chen, Shuo Tong, Yongxin Gu, Tianlong Gong, Chen School of Computer Science and Engineering Nanjing University of Science and Technology China Center for Advanced Intelligence Project RIKEN Japan State Key Laboratory of Complex & Critical Software Environment Beihang University China Jinan University China Department of Automation Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University China

Heterogeneity is a fundamental and challenging issue in federated learning, especially for the graph data due to the complex relationships among the graph nodes. To deal with the heterogeneity, lots of existing methods perform the weighted federation based on their calculated similarities between pairwise clients (i.e., subgraphs). However, their inter-subgraph similarities estimated with the outputs of local models are less reliable, because the final outputs of local models may not comprehensively represent the real distribution of subgraph data. In addition, they ignore the critical intra-heterogeneity which usually exists within each subgraph itself. To address these issues, we propose a novel Federated learning method by integrally modeling the Inter-Intra Heterogeneity (FedIIH). For the inter-subgraph relationship, we propose a novel hierarchical variational model to infer the whole distribution of subgraph data in a multi-level form, so that we can accurately characterize the inter-subgraph similarities with the global perspective. For the intra-heterogeneity, we disentangle the subgraph into multiple latent factors and partition the model parameters into multiple parts, where each part corresponds to a single latent factor. Our FedIIH not only properly computes the distribution similarities between subgraphs, but also learns disentangled representations that are robust to irrelevant factors within subgraphs, so that it successfully considers the inter- and intra- heterogeneity simultaneously. Extensive experiments on six homophilic and five heterophilic graph datasets in both non-overlapping and overlapping settings demonstrate the effectiveness of our method when compared with nine state-of-the-art methods. Specifically, FedIIH averagely outperforms the second-best method by a large margin of 5.79% on all heterophilic datasets. Code is available at https://***/blgpb/FedIIH. Copyright © 2024, The Authors. All rights reserved.

关键词： Federated learning

Mixed one-bit compressive sensing with application to overexposure correction for CT reconstruction

学校读者我要写书评

暂无评论

arXiv 2017年

作者： Huang, Xiaolin Xia, Yan Shi, Lei Huang, Yixing Yan, Ming Hornegger, Joachim Maier, Andreas Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Shanghai China Department of Radiology Stanford University CA United States School of Mathematical Sciences Fudan University Shanghai China Pattern Recognition Lab of Friedrich-Alexander-Universität Erlangen-Nürnberg Erlangen Germany Department of Computational Mathematics Science and Engineering Michigan State University MI United States

When a measurement falls outside the quantization or measurable range, it becomes saturated and cannot be used in classical reconstruction methods. For example, in C-arm angiography systems, which provide projection radiography, fluoroscopy, digital subtraction angiography, and are widely used for medical diagnoses and interventions, the limited dynamic range of C-arm flat detectors leads to overexposure in some projections during an acquisition, such as imaging relatively thin body parts (e.g., the knee). Aiming at overexposure correction for computed tomography (CT) reconstruction, we in this paper propose a mixed one-bit compressive sensing (M1bitCS) to acquire information from both regular and saturated measurements. This method is inspired by the recent progress on one-bit compressive sensing, which deals with only sign observations. Its successful applications imply that information carried by saturated measurements is useful to improve recovery quality. For the proposed M1bit-CS model, alternating direction methods of multipliers is developed and an iterative saturation detection scheme is established. Then we evaluate M1bit-CS on one-dimensional signal recovery tasks. In some experiments, the performance of the proposed algorithms on mixed measurements is almost the same as recovery on unsaturated ones with the same amount of measurements. Finally, we apply the proposed method to overexposure correction for CT reconstruction on a phantom and a simulated clinical image. The results are promising, as the typical streaking artifacts and capping artifacts introduced by saturated projection data are effectively reduced, yielding significant error reduction compared with existing algorithms based on extrapolation. Copyright © 2017, The Authors. All rights reserved.

关键词： Compressed sensing

Hybrid Data-Free Knowledge Distillation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Tang, Jialiang Chen, Shuo Gong, Chen School of Computer Science and Engineering Nanjing University of Science and Technology China Key Laboratory of Intelligent Perception and Systems for High-Dimensional Information of Ministry of Education China Jiangsu Key Laboratory of Image and Video Understanding for Social Security China Center for Advanced Intelligence Project RIKEN Japan Department of Automation Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University China

Data-free knowledge distillation aims to learn a compact student network from a pre-trained large teacher network without using the original training data of the teacher network. Existing collection-based and generation-based methods train student networks by collecting massive real examples and generating synthetic examples, respectively. However, they inevitably become weak in practical scenarios due to the difficulties in gathering or emulating sufficient real-world data. To solve this problem, we propose a novel method called Hybrid Data-Free Distillation (HiDFD), which leverages only a small amount of collected data as well as generates sufficient examples for training student networks. Our HiDFD comprises two primary modules, i.e., the teacher-guided generation and student distillation. The teacher-guided generation module guides a Generative Adversarial Network (GAN) by the teacher network to produce high-quality synthetic examples from very few real-world collected examples. Specifically, we design a feature integration mechanism to prevent the GAN from overfitting and facilitate the reliable representation learning from the teacher network. Meanwhile, we drive a category frequency smoothing technique via the teacher network to balance the generative training of each category. In the student distillation module, we explore a data inflation strategy to properly utilize a blend of real and synthetic data to train the student network via a classifier-sharing-based feature alignment technique. Intensive experiments across multiple benchmarks demonstrate that our HiDFD can achieve state-of-the-art performance using 120 times less collected data than existing methods. Code is available at https://***/tangjialiang97/HiDFD. Copyright © 2024, The Authors. All rights reserved.

关键词： Students

Evaluating deformation patterns of the thoracic aorta in gated CTA sequences

学校读者我要写书评

暂无评论

Evaluating deformation patterns of the thoracic aorta in gat...

IEEE International Symposium on Biomedical Imaging

作者： Ernst Schwartz Roman Gottardi Johannes Holfeld Christian Loewe Martin Czerny Georg Langs Computational Image Analysis and Radiology Lab (CIR) Medical University of Vienna Austria Pattern Recognition and Image Processing Group University of Technology Vienna Austria Division of Cardiothoracic Surgery Department of Surgery Medical University of Vienna Austria Division of Interventional Radiology Department of Radiology Medical University of Vienna Austria Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology USA

ISBN: (纸本)9781424441259;9781424441266

Cardiovascular interventions in the region of the aortic isthmus such as stent-grafting and vessel transposition introduce substantial changes in the deformation properties of the affected vessels. The changes play a fundamental role in the long-term prognosis for any such treatment, but are only poorly understood to date. We explore a fully automated method to quantify the deformation patterns of the thoracic aorta in gated computed tomography sequences. The aorta is segmented by a level set approach that accurately identifies the vessel lumen in each frame of the sequence. Consequently, landmarks on the vessel wall in each frame are registered using a probabilistic method. This allows for the measurement of global and local deformation properties. We evaluate our method on synthetic datasets and report first results of its application on real world data.

关键词： Biomedical imaging image segmentation Radiology Cardiology Surgery Computed tomography Level set Electrocardiography Pathology Deformable models

Dim small targets fusion detection in infrared image

学校读者我要写书评

暂无评论

Dim small targets fusion detection in infrared image

International Conference on Machine Learning and Cybernetics (ICMLC)

作者： Yu-Qiu Sun Yu Zheng Jin-Wen Tian Jian Liu School of Information and Mathematics Yangtze University Jingzhou China State Education Commission Key Laboratory for Image Processing and Intelligent Control Institute for Pattern Recognition and Artificial Intelligence Huazhong University of Science and Technology Wuhan China State Education Commission Key Laboratory for Image Processing and Intelligent Control Department of Electronic Information and Enginery of Huazhong University of Science and Technology Wuhan China

To infrared images, the contrast of target and background is low, dim small targets have no concrete shapes and their textures cannot be reliable predicted. The paper puts forward a novel algorithm to fuse mid-wave and long-wave infrared images and detect targets. Firstly, the source images are decomposed by wavelet transformation. In usual, targets in infrared images are man-made, and their fractal dimension is different comparing with natural background. In wavelet transformation domain high-frequency part, we calculate local fractal dimension and set up fusion rule to merge corresponding sub-images of two matching source images. In low-frequency, we extract local maximum gray level to fuse them. Then reconstruct image by wavelet inverse transformation and obtain fused result image. In fusion results, the contrast between targets and background has obvious changes. And targets can be detected using contrast threshold. The experimental results show that the method proposed in this paper using wavelet transformation fractal dimension to fuse dual band infrared images, and then detect targets is better than using mid-wave or long -wave infrared images detect targets alone.

关键词： Fractals Wavelet transforms Pixel Object detection Fuses Approximation methods Manganese

Background suppression based-on wavelet transformation to detect infrared target

学校读者我要写书评

暂无评论

Background suppression based-on wavelet transformation to de...

International Conference on Machine Learning and Cybernetics (ICMLC)

作者： Yu-Qiu Sun Jin-Wen Tian Jian Liu State Education Commission Key Laboratory for Image Processing and Intelligent Control Institute for Pattern Recognition and Artificial Intelligence Huazhong University of Science and Technology Wuhan China School of Information and Mathematics Yangtze University Jinzhou China State Education Commission Key Laboratory for Image Processing and Intelligent Control Department of Electronic Information and Enginery Huazhong University of Science and Technology Wuhan China

Target detection and location in infrared clutter background is very important to infrared search and track system. Especially for small target detection in infrared image in background of sea and sky, there are no geometric and structure character to use. When targets such as ship and naval vessel sailing at long distance, they always appear around the sea-level line, and it is mixed with cloud and sea-wave clutter. It is difficult to segment and locate precisely. Background suppression based-on wavelet transformation is proposed in the paper. Wavelet decomposition makes it possible to analyze a signal both in time and frequency domains. In the paper infrared sea and key background images are processed. For their SNR is low and background is complex, using wavelet transformation decomposes an original image and extract approximate feature to reconstruct an image, which mainly includes background information. A new image would be obtained using background image subtracted from original image. There are mainly target and noise points left in the new image. By setting proper threshold, the target can be detected perfectly. At last, experiment results are given and show the method is practical.

关键词： Infrared detectors Object detection Infrared imaging Target tracking Marine vehicles Clouds image segmentation Wavelet analysis Wavelet domain Signal analysis

A regularization approach for instance-based superset label learning

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Gong, Chen Liu, Tongliang Tang, Yuanyan Yang, Jian Yang, Jie Tao, Dacheng School of Computer Science and Engineering Nanjing University of Science and Technology Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University School of Software Faculty of Engineering and Information Technology University of Technology Sydney UltimoNSW2007 Australia Faculty of Science and Technology University of Macau Macau999078 China College of Computer Science Chongqing University Chongqing400000 China School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing210094 China Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University Shanghai200240 China School of Information Technologies Faculty of Engineering and Information Technologies University of Sydney J12/318 Cleveland St DarlingtonNSW2008 Australia

Different from the traditional supervised learning in which each training example has only one explicit label, Superset Label Learning (SLL) refers to the problem that a training example can be associated with a set of candidate labels, and only one of them is correct. Existing SLL methods are either regularization-based or instance-based, and the latter of which has achieved state-of-the-art performance. This is because the latest instance-based methods contain an explicit disambiguation operation that accurately picks up the groundtruth label of each training example from its ambiguous candidate labels. However, such disambiguation operation does not fully consider the mutually exclusive relationship among different candidate labels, so the disambiguated labels are usually generated in a nondiscriminative way, which is unfavorable for the instance-based methods to obtain satisfactory performance. To address this defect, we develop a novel Regularization approach for Instancebased Superset Label (RegISL) learning so that our instancebased method also inherits the good discriminative ability possessed by the regularization scheme. Specifically, we employ a graph to represent the training set, and require the examples that are adjacent on the graph to obtain similar labels. More importantly, a discrimination term is proposed to enlarge the gap of values between possible labels and unlikely labels for every training example. As a result, the intrinsic constraints among different candidate labels are deployed, and the disambiguated labels generated by RegISL are more discriminative and accurate than those output by existing instance-based algorithms. The experimental results on various tasks convincingly demonstrate the superiority of our RegISL to other typical SLL methods in terms of both training accuracy and test accuracy. Copyright © 2019, The Authors. All rights reserved.

关键词： Machine learning

Partial Differential Equations is All You Need for Generating Neural Architectures - A Theory for Physical Artificial Intelligence Systems

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Guo, Ping Huang, Kaizhu Xu, Zenglin Image Processing & Pattern Recognition Lab. Beijing Normal University Beijing100875 China Data Science Research Center Duke Kunshan University Jiangsu Kunshan215316 China School of Computer Science and Technology Harbin Institute of Technology at ShenZhen Peng Cheng National Lab Guangdong Shenzhen510855 China

In this work, we generalize the reaction-diffusion equation in statistical physics, Schrödinger equation in quantum mechanics, and Helmholtz equation in paraxial optics into the neural partial differential equations (NPDE), which can be considered as the fundamental equations in the field of artificial intelligence research. We take the finite difference method to discretize NPDEs for finding the numerical solution. Moreover, the basic building blocks of deep neural network architectures, including multi-layer perceptron, convolutional neural networks, and recurrent neural networks, are generated. The learning strategies, such as Adaptive moment estimation, L-BFGS, pseudoinverse learning algorithms, and partial differential equation constrained optimization, are also developed. This work is of significance in that it presents a clear view of interpretable deep neural networks physically, enables high possibility to be applied to analog computing device design, and paves the road to physical artificial intelligence. © 2021, CC BY-NC-SA.

关键词： Schrodinger equation

Fast signal recovery from saturated measurements by linear loss and nonconvex penalties

学校读者我要写书评

暂无评论

arXiv 2018年

作者： He, Fan Huang, Xiaolin Liu, Yipeng Yan, Ming Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University The MOE Key Laboratory of System Control and Information Processing Shanghai200240 China School of Information and Communication Engineering University of Electronic Science and Technology of China Chengdu611731 China The Department of Computational Mathematics Science and Engineering Michigan State University MI United States

Sign information is the key to overcoming the inevitable saturation error in compressive sensing systems, which causes information loss and results in bias. For sparse signal recovery from saturation, we propose to use a linear loss to improve the effectiveness from existing methods that utilize hard constraints/hinge loss for sign consistency. Due to the use of linear loss, an analytical solution in the update progress is obtained, and some nonconvex penalties are applicable, e.g., the minimax concave penalty, the 0 norm, and the sorted 1 norm. Theoretical analysis reveals that the estimation error can still be bounded. Generally, with linear loss and nonconvex penalties, the recovery performance is significantly improved, and the computational time is largely saved, which is verified by the numerical experiments. Copyright © 2018, The Authors. All rights reserved.

关键词： Compressed sensing