检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Pan, Qingtao Qiao, Wenhao Lou, Jingjiao Ji, Bing Li, Shuo School of Control Science and Engineering Shandong University Jinan China Key Laboratory of Machine Intelligence and System Control Ministry of Education China Department of Computer and Data Science Department of Biomedical Engineering Case Western Reserve University United States

Semi-supervised medical image segmentation (SSMIS) uses consistency learning to regularize model training, which alleviates the burden of pixel-wise manual annotations. However, it often suffers from error supervision from low-quality pseudo labels. Vision-Language Model (VLM) has great potential to enhance pseudo labels by introducing text prompt guided multimodal supervision information. It nevertheless faces the cross-modal problem: the obtained messages tend to correspond to multiple targets. To address aforementioned problems, we propose a Dual Semantic Similarity-Supervised VLM (DuSSS) for SSMIS. Specifically, 1) a Dual Contrastive Learning (DCL) is designed to improve cross-modal semantic consistency by capturing intrinsic representations within each modality and semantic correlations across modalities. 2) To encourage the learning of multiple semantic correspondences, a Semantic Similarity-Supervision strategy (SSS) is proposed and injected into each contrastive learning process in DCL, supervising semantic similarity via the distribution-based uncertainty levels. Furthermore, a novel VLM-based SSMIS network is designed to compensate for the quality deficiencies of pseudo-labels. It utilizes the pretrained VLM to generate text prompt guided supervision information, refining the pseudo label for better consistency regularization. Experimental results demonstrate that our DuSSS achieves outstanding performance with Dice of 82.52%, 74.61% and 78.03% on three public datasets (QaTa-COV19, BM-Seg and MoNuSeg). Copyright © 2024, The Authors. All rights reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Domain-Guided Conditional Diffusion Model for Unsupervised Domain Adaptation

arXiv

引用

arXiv 2023年

作者： Zhang, Yulong Chen, Shuhao Jiang, Weisen Zhang, Yu Lu, Jiangang Kwok, James T. The State Key Laboratory of Industrial Control Technology College of Control Science and Engineering Zhejiang University Hangzhou China The Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen China The Department of Computer Science and Engineering Southern University of Science and Technology Hong Kong University of Science and Technology Hong Kong The Zhejiang Laboratory Hangzhou China The Department of Computer Science and Engineering Hong Kong University of Science and Technology Hong Kong

Limited transferability hinders the performance of deep learning models when applied to new application scenarios. Recently, Unsupervised Domain Adaptation (UDA) has achieved significant progress in addressing this issue via learning domain-invariant features. However, the performance of existing UDA methods is constrained by the large domain shift and limited target domain data. To alleviate these issues, we propose DomAin-guided Conditional Diffusion Model (DACDM) to generate high-fidelity and diversity samples for the target domain. In the proposed DACDM, by introducing class information, the labels of generated samples can be controlled, and a domain classifier is further introduced in DACDM to guide the generated samples for the target domain. The generated samples help existing UDA methods transfer from the source domain to the target domain more easily, thus improving the transfer performance. Extensive experiments on various benchmarks demonstrate that DACDM brings a large improvement to the performance of existing UDA methods. Copyright © 2023, The Authors. All rights reserved.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Continuous-Time Damping-Based Mirror Descent for a Class of Non-Convex Multi-Player Games with Coupling Constraints

Continuous-Time Damping-Based Mirror Descent for a Class of ...

引用

International Conference on control and Automation (ICCA)

作者： Guanpu Chen Kun Cao Karl Henrik Johansson Yiguang Hong Division of Decision and Control Systems School of Electrical Engineering and Computer Science KTH Royal Institute of Technology Digital Futures Stockholm Sweden Department of Control Science and Engineering Research Institute for Intelligent Autonomous Systems Tongji University Shanghai China

ISBN: (数字)9798350354409

ISBN: (纸本)9798350354416

We study the computation of the global generalized Nash equilibrium (GNE) for a class of non-convex multi-player games, where players' actions are subject to both local and coupling constraints. Due to the non-convex payoff functions, we employ canonical duality to reformulate the setting as a complementary problem. Under given conditions, we reveal the relation between the stationary point and the global GNE. According to the convex-concave properties within the complementary function, we propose a continuous-time mirror descent to compute GNE by generating functions in the Bregman divergence and the damping-based design. Then, we devise several Lyapunov functions to prove that the trajectory along the dynamics is bounded and convergent.

关键词： Couplings Damping Automation Games Nash equilibrium Trajectory Mirrors

来源：评论

学校读者我要写书评

暂无评论

Proposal of a novel peeling of nano-particle (PNP) process for localized material removal on a hard material surface by controllable magnetic fields 22

Proposal of a novel peeling of nano-particle (PNP) process f...

引用

22nd International Conference of the European Society for Precision engineering and Nanotechnology, EUSPEN 2022

作者： Permpatdechakul, Thitipat Khajornrungruang, Panart Suzuki, Keisuke Kutomi, Shotaro Graduate School of Computer Science and System Engineering Mechanical Information Science and Technology Course Kyushu Institute of Technology Fukuoka Japan Department of Intelligent and Control Systems Avanced Mechanical Division Kyushu Institute of Technology Fukuoka Japan School of Computer Science and System Engineering Advanced Mechanical Course Kyushu Institute of Technology Fukuoka Japan

ISBN: (纸本)9781998999118

This article proposes a novel Peeling of Nano-Particle (PNP) process to locally remove material on a hard material surface using controllable magnetic fields. Fe3O4 particles in the size range of 50-100 nm in aqueous solution (pH value 11 adjusted by KOH) were magnetically controlled to adhere and remove from a 4H-SiC surface by two electromagnetic coils, which were sandwiched between the substrate. The particle controlling phenomenon on the surface was experimentally observed and verified by our developed apparatus applying an evanescent field. According to the results, particles could be controlled to approach and remove from the substrate surface with 12mT of magnetic field at the selective area on the substrate surface to realise the proposed PNP process. © European Society for Precision engineering and Nanotechnology, Conference Proceedings - 22nd International Conference and Exhibition, EUSPEN 2022. All rights reserved.

关键词： Magnetic fields

来源：评论

学校读者我要写书评

暂无评论

Distributed Online Bandit Nonconvex Optimization with One-Point Residual Feedback via Dynamic Regret

arXiv

引用

arXiv 2024年

作者： Hua, Youqing Liu, Shuai Hong, Yiguang Johansson, Karl Henrik Wang, Guangchen School of Control Science and Engineering Shandong University Jinan China Nanyang Technological University Singapore Department of Control Science and Engineering Shanghai Research Institute for Intelligent Autonomous Systems Tongji University Shanghai China School of Electrical Engineering and Computer Science KTH Royal Institute of Technology Stockholm Sweden

This paper considers the distributed online bandit optimization problem with nonconvex loss functions over a time-varying digraph. This problem can be viewed as a repeated game between a group of online players and an adversary. At each round, each player selects a decision from the constraint set, and then the adversary assigns an arbitrary, possibly nonconvex, loss function to this player. Only the loss value at the current round, rather than the entire loss function or any other information (e.g. gradient), is privately revealed to the player. Players aim to minimize a sequence of global loss functions, which are the sum of local losses. We observe that traditional multi-point bandit algorithms are unsuitable for online optimization, where the data for the loss function are not all a priori, while the one-point bandit algorithms suffer from poor regret guarantees. To address these issues, we propose a novel one-point residual feedback distributed online algorithm. This algorithm estimates the gradient using residuals from two points, effectively reducing the regret bound while maintaining O(1) sampling complexity per iteration. We employ a rigorous metric, dynamic regret, to evaluate the algorithm's performance. By appropriately selecting the step size and smoothing parameters, we demonstrate that the expected dynamic regret of our algorithm is comparable to existing algorithms that use two-point feedback, provided the deviation in the objective function sequence and the path length of the minimization grows sublinearly. Finally, we validate the effectiveness of the proposed algorithm through numerical simulations. Copyright © 2024, The Authors. All rights reserved.

关键词： Optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Paper Defects Classification Based on VGG16 and Transfer Learning

引用

Palpu Chongi Gisul/Journal of Korea Technical Association of the Pulp and Paper Industry 2021年第2期53卷 5-14页

作者： Yun-Hui, Qu Tang, Wei Feng, Bo Computer Teaching and Research Section Xi’an Medical University Xi’an Shaanxi710021 China Department of Electric and Control Engineering Shaanxi University of Science & Technology Xi’an Shaanxi710021 China Department of Electric and Control Engineering Shaanxi University of Science & Technology Xi’an Shaanxi710021 China

There are some problems in traditional paper defects classification, such as the poor generalization performance, less types of recognition, and insufficient recognition accuracy. The deep learning method provides a new scheme for paper defects classification. However, due to the small sample size of paper defect images set, the over fitting phenomenon is easy to appear in the training process. Aiming this problem, a transfer learning method based on convolutional neural network model is ***, freezing the first seven construction layers of VGG16 network which has been trained by ImageNet, and fine tune the rest convolution layers with the paper defect images set to complete the feature extraction;Secondly, the full connection layers for classification are improved to meet the needs of paper defects classification;Finally, transfer learning strategy is adopted in the training process to improve the efficiency. The experimental results demonstrate that the paper defects classification proposed in our approach can improve the efficiency and accuracy of paper defects recognition. The approach will beneficial for the web inspection process. © 2021 Korean Technical Assoc. of the Pulp and Paper Industry. All rights reserved.

关键词： Efficiency

来源：评论

学校读者我要写书评

暂无评论

Reversible Data Hiding Scheme Based on Prediction Error Bit-plane Compression

Reversible Data Hiding Scheme Based on Prediction Error Bit-...

引用

IEEE International Conference on Big Data and Cloud Computing (BdCloud)

作者： Guo Yuan Chao Heng Zhai Ping Liu Ziqi Li Haibo Li Wenpeng College of computer science and technology Heilongjiang University Harbin China College of Computer and Control Engineering Qiqihar University Qiqihar China

How to maximize embedding capacity is one of the current challenges in the field of reversible data hiding. A reversible data hiding scheme is proposed based on the rearrangement and compression of prediction error bit-planes in this paper. The image holder first predicts the pixels to obtain the error map, then decomposes them into bit-planes, rearranges and compresses them to maximize the embedding space, and embeds the auxiliary information and secret information into the freed space according to certain rules to generate a cryptographic image. After acquiring the cryptographic image, the image receiver extracts the embedded secret information and recovers the original image using the auxiliary information. In this paper, a rearrangement strategy based on full pixel correlation was built to maximize compression efficiency and further compress the auxiliary information to free up more embeddable space and improve embedding capacity. A large number of test has revealed that the embedding rate of this method is 30% higher on average than the current state-of-the-art algorithm.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Contracting Skeletal Kinematics for Human-Related Video Anomaly Detection

arXiv

引用

arXiv 2023年

作者： Flaborea, Alessandro di Melendugno, Guido Maria D'Amely D'Arrigo, Stefano Sterpa, Marco Aurelio Sampieri, Alessio Galasso, Fabio Department of Computer Science Sapienza University of Rome Italy Department of Computer Control and Management Engineering Sapienza University of Rome Italy

Detecting the anomaly of human behavior is paramount to timely recognizing endangering situations, such as street fights or elderly falls. However, anomaly detection is complex since anomalous events are rare and because it is an open set recognition task, i.e., what is anomalous at inference has not been observed at training. We propose COSKAD, a novel model that encodes skeletal human motion by a graph convolutional network and learns to COntract SKeletal kinematic embeddings onto a latent hypersphere of minimum volume for Video Anomaly Detection. We propose three latent spaces: the commonlyadopted Euclidean and the novel spherical and hyperbolic. All variants outperform the state-of-the-art on the most recent UBnormal dataset, for which we contribute a human-related version with annotated skeletons. COSKAD sets a new state-of-the-art on the human-related versions of ShanghaiTech Campus and CUHK Avenue, with performance comparable to video-based methods. Source code and dataset will be released upon acceptance. Copyright © 2023, The Authors. All rights reserved.

关键词： Graph embeddings

来源：评论

学校读者我要写书评

暂无评论

Cluster Synchronization of Kuramoto Oscillators via Pacemakers and Mean-phase Feedback control 63

Cluster Synchronization of Kuramoto Oscillators via Pacemake...

引用

63rd IEEE Conference on Decision and control, CDC 2024

作者： Kokubo, Ryota Matsubara, Mitsuaki Kato, Rui Ishii, Hideaki Tokyo Institute of Technology Department of Computer Science Yokohama226-8502 Japan Tokyo Institute of Technology Department of Systems and Control Engineering Yokohama226-8502 Japan Villeurbanne69622 France The University of Tokyo Department of Information Physics and Computing Tokyo113-8656 Japan

ISBN: (纸本)9798350316339

Brain networks typically exhibit characteristic synchronization patterns where several synchronized clusters coexist. On the other hand, neurological disorders are considered to be related to pathological synchronization such as excessive synchronization of large populations of neurons. Motivated by these phenomena, this paper presents two approaches to control the cluster synchronization of Kuramoto oscillators. One is based on the use of pacemakers to the clusters, and the other is based on feeding back the mean phases to the clusters. We first show conditions on the pacemaker weights and the feedback gains for the network to achieve cluster synchronization. Then, we propose a method to find optimal feedback gains through convex optimization. A numerical example demonstrates the effectiveness of the proposed methods. © 2024 IEEE.

关键词： Pacemakers

来源：评论

学校读者我要写书评

暂无评论

Incremental High Fuzzy Utility Itemset Mining 22

Incremental High Fuzzy Utility Itemset Mining

引用

9th Multidisciplinary International Social Networks Conference, MISNC 2022

作者： Hong, Tzung-Pei Hung, Wei-Teng Huang, Wei-Ming Tsai, Yu-Chuan Department of Computer Science and Information Engineering National University of Kaohsiung Taiwan Department of Computer Science and Engineering National Sun Yat-Sen University Taiwan Department of Electrical and Control China Steel Inc. Taiwan Library and Information Center National University of Kaohsiung Kaohsiung811 Taiwan

ISBN: (纸本)9781450398435

In data mining, frequent-pattern mining methods are used for handling binary databases. Utility mining addresses this limitation by considering the item utilities and quantities when discovering the high utility itemsets. In addition, to make these high utility patterns more human-explainable, we use the fuzzy-set theory to the utility mining algorithms. However, real-world databases are usually dynamic. That is, new transactions may be intermittently added, and the corresponding mined knowledge needs to be updated. This paper uses the famous incremental strategy, fast update (FUP), to modify high fuzzy utility itemsets from new coming data. We implement the FUP strategy on the Apriori-based approach for maintaining up-to-date high fuzzy utility *** also conducted experiments to demonstrate that the incremental algorithm significantly outperforms the Apriori-based batch mining method. © 2022 ACM.

关键词： Data mining

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：