检索结果-内蒙古大学图书馆

38th Conference on Neural Information Processing systems, NeurIPS 2024

作者： Jena, Rohit Sethi, Deeksha Chaudhari, Pratik Gee, James C. Computer and Information Science United States Electrical and Systems Engineering United States Radiology United States Penn Image Computing and Science Laboratory United States

Classical optimization and learning-based methods are the two reigning paradigms in deformable image registration. While optimization-based methods boast generalizability across modalities and robust performance, learning-based methods promise peak performance, incorporating weak supervision and amortized optimization. However, the exact conditions for either paradigm to perform well over the other are shrouded and not explicitly outlined in the existing literature. In this paper, we make an explicit correspondence between the mutual information of the distribution of per-pixel intensity and labels, and the performance of classical registration methods. This strong correlation hints to the fact that architectural designs in learning-based methods is unlikely to affect this correlation, and therefore, the performance of learning-based methods. This hypothesis is thoroughly validated with state-of-the-art classical and learning-based methods. However, learning-based methods with weak supervision can perform high-fidelity intensity and label registration, which is not possible with classical methods. Next, we show that this high-fidelity feature learning does not translate to invariance to domain shift, and learning-based methods are sensitive to such changes in the data distribution. We reassess and recalibrate performance expectations from classical and DLIR methods under access to label supervision, training time, and its generalization capabilities under minor domain shifts. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Learning discriminative foreground-and-background features for few-shot segmentation

引用

Multimedia Tools and Applications 2024年第18期83卷 55999-56019页

作者： Jiang, Cong Zhou, Yange Liu, Zhaoshuo Feng, Chaolu Li, Wei Yang, Jinzhu School of Computer Science and Engineering Northeastern University Liaoning Shenyang110819 China College of Science Northeastern University Liaoning Shenyang110819 China Key Laboratory of Intelligent Computing in Medical Image Ministry of Education Liaoning Shenyang110819 China National Frontiers Science Center for Industrial Intelligence and Systems Optimization Liaoning Shenyang110819 China

Few-shot Semantic Segmentation (FSS) endeavors to segment novel categories in a query image by referring to a support set comprising only a few annotated examples. Presently, many existing FSS methodologies primarily embrace the prototype learning paradigm and concentrate on optimizing the matching mechanism. However, these approaches tend to overlook the discrimination between the features of foreground background. Consequently, the segmentation results are often imprecise when it comes to capturing intricate structures, such as boundaries and small objects. In this study, we introduce the Discriminative Foreground-and-Background feature learning Network (DFBNet) to enhance the distinguishability of bilateral features. DFBNet comprises three major modules: a multi-level self-matching module (MSM), a feature separation module (FSM), and a semantic alignment module (SAM). The MSM generates prior masks separately for the foreground and background, employing a self-matching strategy across different feature levels. These prior masks are subsequently used as scaling factors within the FSM, where the features of the query’s foreground and background are independently scaled up and then concatenated along the channel dimension. Furthermore, we incorporate a two-layer Transformer encoder-based semantic alignment module (SAM) in DFBNet to refine the features, thereby creating a greater distinction between the foreground and background features. The performance of DFBNet is evaluated on the PASCAL-5i and COCO-20i benchmarks, demonstrating its superiority over existing solutions and establishing new state-of-the-art results in the field of few-shot semantic segmentation. The codes will be released if this paper is accepted. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2023.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Deep Learning in Medical image Registration: Magic or Mirage?

arXiv

引用

arXiv 2024年

作者： Jena, Rohit Sethi, Deeksha Chaudhari, Pratik Gee, James C. Computer and Information Science Electrical and Systems Engineering Radiology Penn Image Computing and Science Laboratory United States

Classical optimization and learning-based methods are the two reigning paradigms in deformable image registration. While optimization-based methods boast generalizability across modalities and robust performance, learning-based methods promise peak performance, incorporating weak supervision and amortized optimization. However, the exact conditions for either paradigm to perform well over the other are shrouded and not explicitly outlined in the existing literature. In this paper, we make an explicit correspondence between the mutual information of the distribution of per-pixel intensity and labels, and the performance of classical registration methods. This strong correlation hints to the fact that architectural designs in learning-based methods is unlikely to affect this correlation, and therefore, the performance of learning-based methods. This hypothesis is thoroughly validated with state-of-the-art classical and learning-based methods. However, learning-based methods with weak supervision can perform high-fidelity intensity and label registration, which is not possible with classical methods. Next, we show that this high-fidelity feature learning does not translate to invariance to domain shift, and learning-based methods are sensitive to such changes in the data distribution. Finally, we propose a general recipe to choose the best paradigm for a given registration problem, based on these observations. © 2024, CC BY.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

BECNN: Bias Field Estimation CNN Trained with a Dual Route Implicit Supervised Learning Strategy

BECNN: Bias Field Estimation CNN Trained with a Dual Route I...

引用

2024 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2024

作者： Chen, Shuaizheng Feng, Chaolu Li, Wei Yang, Jinzhu Zhao, Dazhe Northeastern University School of Computer Science And Engineering Shenyang China Ministry of Education Northeastern University Key Laboratory of Intelligent Computing in Medical Image Shenyang China Northeastern University National Frontiers Science Center for Industrial Intelligence and Systems Optimization Shenyang China

ISBN: (纸本)9798350386226

Bias fields adversely affect various automatic analysis technologies. Therefore, bias field correction is essential. However, deep learning based methods encounter challenges in obtaining ground truth. Although existing methods attempt to address this problem by using training-free techniques or constructing datasets with approximations of the ground truth, the lack of task-oriented guidance, training instability, and inappropriate use of approximations still impact performance. Additionally, different approaches for bias field correction, i.e., estimating bias fields versus directly restoring clean images, exhibit different performances. However, there is no consensus on the best way, resulting in limited performance in some cases. To address these problems, we propose a bias field generation method to construct the dataset and provide task-oriented information. We then propose the concept of Equivalent of Residual Mapping (ERM) to analyze the advantages of estimating bias fields. According to ERM, we propose the Bias field Estimation Convolutional Neural Network (BECNN). Finally, we propose the Dual Route Implicit Supervised Learning (DRISL) strategy to balance the guidance derived from approximations of the ground truth with over-dependence on them. The proposed method is compared qualitatively and quantitatively with the correlated methods. Experiment results demonstrate that the proposed method performs effectively both on bias field estimation and correction. © 2024 IEEE.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

ECA-RetinaNet: A Novel Self-Attention RetinaNet for Environmental Microorganism image Object Detection

ECA-RetinaNet: A Novel Self-Attention RetinaNet for Environm...

引用

2023 IEEE International Conference on Big Data, BigData 2023

作者： Yang, Hechen Yang, Jinzhu Jiang, Tao Zhao, Xin Zhang, Jinghua Zhao, Peng Chen, Ao Nie, Qianqing Grzegorzek, Marcin Li, Chen Northeastern University Microscopic Image and Medical Image Analysis Group College of Medicine and Biological Information Engineering Shenyang China Northeastern University Key Laboratory of Intelligent Computing in Medical Image Ministry of Education Liaoning Shenyang China Chengdu University of Traditional Chinese Medicine School of Intelligent Medicine Chengdu China Chengdu University of Information Technology International Joint Institute of Robotics and Intelligent Systems Chengdu China Institute of Medical Informatics University of Luebeck Luebeck Germany

ISBN: (纸本)9798350324457

The detection of environmental microorganisms is always a difficult task, e specially when the multi-scale environment is complex. For tiny objects in microscopic images, current detection methods face the challenge of accurate identification and localization. In contrast, we propose a convolutional neural network (ECA-RetinaNet) for microscopic object detection of which underlying dataset is a high-quality EMDS-7 dataset. The accuracy of ECA-RetinaNet is high, with a high mean Average Precision (mAP) value of 81.42% in the Environmental Microorganisms (EMS) detection task. Its accuracy has been higher than that of the two-stage object detection network. © 2023 IEEE.

关键词： Attention Convolutional Neural Network Environmental Miroorganisms Feature Fusion Object Detection

来源：评论

学校读者我要写书评

暂无评论

SETTP: Style Extraction and Tunable Inference via Dual-Level Transferable Prompt Learning 27

SETTP: Style Extraction and Tunable Inference via Dual-Level...

引用

27th European Conference on Artificial Intelligence, ECAI 2024

作者： Jin, Chunzhen Huang, Yongfeng Wang, Yaqi Cao, Peng Zaïane, Osmar Northeastern University Shenyang China Key Laboratory of Intelligent Computing in Medical Image of Ministry of Education Northeastern University Shenyang China Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong National Frontiers Science Center for Industrial Intelligence and Systems Optimization Shenyang China Amii University of Alberta EdmontonAB Canada

ISBN: (纸本)9781643685489

Text style transfer, an important research direction in natural language processing, aims to adapt the text to various preferences but often faces challenges with limited resources. In this work, we introduce a novel method termed Style Extraction and Tunable Inference via Dual-level Transferable Prompt Learning (SETTP) for effective style transfer in low-resource scenarios. First, SETTP learns source style-level prompts containing fundamental style characteristics from high-resource style transfer. During training, the source style-level prompts are transferred through an attention module to derive a target style-level prompt for beneficial knowledge provision in low-resource style transfer. Additionally, we propose instance-level prompts obtained by clustering the target resources based on the semantic content to reduce semantic bias. We also propose an automated evaluation approach of style similarity based on alignment with human evaluations using ChatGPT-4. Our experiments across three resourceful styles show that SETTP requires only 1/20th of the data volume to achieve performance comparable to state-of-the-art methods. In tasks involving scarce data like writing style and role style, SETTP outperforms previous methods by 16.24%. © 2024 The Authors.

关键词：

来源：评论

学校读者我要写书评

暂无评论

csl-MTFL: Multi-task Feature Learning with Joint Correlation Structure Learning for Alzheimer’s Disease Cognitive Performance Prediction 9th

csl-MTFL: Multi-task Feature Learning with Joint Correlatio...

引用

19th International Conference on Advanced Data Mining and Applications, ADMA 2023

作者： Liang, Wei Zhang, Kai Cao, Peng Liu, Xiaoli Yang, Jinzhu Zaiane, Osmar R. Computer Science and Engineering Northeastern University Shenyang China Key Laboratory of Intelligent Computing in Medical Image of Ministry of Education Northeastern University Shenyang China National Frontiers Science Center for Industrial Intelligence and Systems Optimization Shenyang110819 China DAMO Academy Alibaba Group Hangzhou China Alberta Machine Intelligence Institute University of Alberta EdmontonAB Canada

ISBN: (纸本)9783031466700

Alzheimer’s disease (AD) is a common chronic neurodegenerative disease and the accurate prediction of the clinical cognitive performance is important for diagnosis and treatment. Recently, multi-task feature learning (MTFL) methods with sparsity-inducing regularization have been widely investigated on cognitive performance prediction tasks. Although they have proved to achieve improved performance compared with single-task learning, the major challenges are still not fully resolved. They involve how to capture the non-linear correlation among the tasks or features, and how to introduce the learned correlation for guiding the MTFL learning. To resolve these challenges, we introduce a correlation structure learning method through self-attention learning and sequence learning for jointly capturing the complicated but more flexible relationship for features and tasks, respectively. Moreover, we develop a dual graph regularization to encode the inherent correlation and an efficient optimization algorithm for solving the nonsmooth objective function. Extensive results on the ADNI dataset demonstrate that the proposed joint training framework outperforms existing methods and achieves state-of-the-art prediction performance of AD. Specifically, the proposed algorithm achieves an nMSE (normalized Mean Squared Error)/wR (weighted R-value) of 3.808/0.438, obtaining a relative improvement of 3.84 % /7.35 % compared with the MTFL method. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

关键词： Neurodegenerative diseases

来源：评论

学校读者我要写书评

暂无评论

Towards Time-Variant-Aware Link Prediction in Dynamic Graph Through Self-supervised Learning 9th

Towards Time-Variant-Aware Link Prediction in Dynamic Graph...

引用

19th International Conference on Advanced Data Mining and Applications, ADMA 2023

作者： Wen, Guangqi Cao, Peng Jin, Zhiyong Song, Ruoxian Liu, Xiaoli Yang, Jinzhu Zaiane, Osmar R. Computer Science and Engineering Northeastern University Shenyang China Key Laboratory of Intelligent Computing in Medical Image of Ministry of Education Northeastern University Shenyang China National Frontiers Science Center for Industrial Intelligence and Systems Optimization Shenyang110819 China DAMO Academy Alibaba Group Hangzhou China Alberta Machine Intelligence Institute University of Alberta EdmontonAB Canada

ISBN: (纸本)9783031466731

Dynamic graph link prediction is a challenging problem because the graph topology and node attributes vary at different times. A purely supervised learning scheme for the dynamic graph data usually leads to poor generalization due to insufficient supervision. As a promising solution, self-supervised learning can be introduced to dynamic graph analysis tasks. However, the self-supervised learning paradigm for dynamic graph learning has not been sufficiently investigated due to the complicated properties of the evolving graphs. We assume that the dynamic graphs consist of three independent types of key factors, i.e., graph time-variant information, time-invariant information and noise. Based on this assumption, we propose the Self-supervised Decoupling for Dynamic Graph (SDDG) framework for explicitly decoupling the representation which characterizes these three factors, thus enhancing the interpretability of the learned representation and link prediction performance. More specifically, we design an encoder-decoder architecture to sufficiently exploit the dynamic graph itself with multiple regularizations, so that the time-variant embedding of dynamic graph data can be effectively decoupled from the perspectives of node and structure for time-variant-aware link prediction. Experiments on five benchmark link prediction tasks demonstrate the significant improvement of our SDDG over the state-of-the-art methods. For example, SDDG achieves 98.2% and 97.3% Top-1 AUC on Reddit-B and Reddit-T, outperforming the DDGCL model, by 1.9% and 1.8%, respectively. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

关键词： Supervised learning

来源：评论

学校读者我要写书评

暂无评论

A Novel Federated Learning Based Intrusion Detection System for IoT Networks

A Novel Federated Learning Based Intrusion Detection System ...

引用

IEEE International Conference on Communications (ICC)

作者： Rabaie Benameur Amine Dahane Sami Souihi Abdelhamid Mellouk Laboratory of Industrial Computing and Networks (RIIR) University of Oran 1 Oran Algeria Institute of Applied Science and Technology ISTA Oran Algeria Image Signal and Intelligent Systems (LiSSi) Laboratory TincNET Research Team Univ. Paris-Est Cr eteil France

ISBN: (数字)9781728190549

ISBN: (纸本)9781728190556

In the realm of IoT platforms, susceptibility to cyber-attacks is a pressing concern, necessitating the deployment of Intrusion Detection systems (IDS). Constructing a scalable, accurate, and lightweight model without compromising data privacy poses a formidable challenge. This study assesses classical and novel approaches employing federated learning (FL) to train IDS models. Optimization through Knowledge Distillation (KD) techniques aims to enhance computational efficiency. Experimental results reveal the efficacy of federated learning, achieving an 84.5% accuracy for 15 attack types, and an impressive performance for binary network attack classification. Notably, these models exhibit shorter inference times compared to cutting-edge machine learning models trained on the Edge-IIoTset dataset, offering promising advancements in IoT security.

关键词： Data privacy Accuracy Federated learning Computational modeling image edge detection Intrusion detection Pressing

来源：评论

学校读者我要写书评

暂无评论

Label Correlation Guided Feature Selection for Multi-label Learning 9th

Label Correlation Guided Feature Selection for Multi-label ...

引用

19th International Conference on Advanced Data Mining and Applications, ADMA 2023

作者： Zhang, Kai Liang, Wei Cao, Peng Yang, Jinzhu Li, Weiping Zaiane, Osmar R. Computer Science and Engineering Northeastern University Shenyang China Key Laboratory of Intelligent Computing in Medical Image of Ministry of Education Northeastern University Shenyang China National Frontiers Science Center for Industrial Intelligence and Systems Optimization Shenyang110819 China School of Software and Microelectronics Peking University Beijing China Alberta Machine Intelligence Institute University of Alberta EdmontonAB Canada

ISBN: (纸本)9783031466731

Multi-label learning has received much attention due to its wide range of application domains. Multi-label data often has high-dimensional features, which brings more challenges to classification algorithms. Feature selection based on sparse learning is one of the most important approaches, which can select discriminative features to alleviate the curse of dimensionality. However, most of these methods do not consider feature redundancy and label correlation simultaneously. In this work, we propose a multi-label feature selection method that takes feature redundancy and label correlation into account. Specifically, the proposed method first divides features into groups according to feature correlation and then encourages competition within each group but relaxes competition between groups to eliminate redundant and irrelevant features. Moreover, we propose an approach to capture label correlation and exploit it to improve the feature selection. Finally, an iterative optimization algorithm is designed to obtain the feature weights for multi-label feature selection. Extensive experiments on various multi-label datasets demonstrate the superiority of the proposed method compared with some state-of-the-art multi-label feature selection methods. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

关键词： Feature Selection

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：