检索结果-内蒙古大学图书馆

GPS: graph contrastive learning via multi-scale augmented views from adversarial pooling

Science China(information Sciences) 2025年第1期68卷 145-158页

作者： Wei JU Yiyang GU Zhengyang MAO Ziyue QIAO Yifang QIN Xiao LUO Hui XIONG Ming ZHANG School of Computer Science National Key Laboratory for Multimedia Information ProcessingPeking University Artificial Intelligence Thrust The Hong Kong University of Science and Technology Department of Computer Science University of California

Self-supervised graph representation learning has recently shown considerable promise in a range of fields, including bioinformatics and social networks. A large number of graph contrastive learning approaches have shown promising performance for representation learning on graphs, which train models by maximizing agreement between original graphs and their augmented views(i.e., positive views). Unfortunately, these methods usually involve pre-defined augmentation strategies based on the knowledge of human experts. Moreover, these strategies may fail to generate challenging positive views to provide sufficient supervision signals. In this paper, we present a novel approach named graph pooling contrast(GPS) to address these *** by the fact that graph pooling can adaptively coarsen the graph with the removal of redundancy, we rethink graph pooling and leverage it to automatically generate multi-scale positive views with varying emphasis on providing challenging positives and preserving semantics, i.e., strongly-augmented view and weakly-augmented view. Then, we incorporate both views into a joint contrastive learning framework with similarity learning and consistency learning, where our pooling module is adversarially trained with respect to the encoder for adversarial robustness. Experiments on twelve datasets on both graph classification and transfer learning tasks verify the superiority of the proposed method over its counterparts.

关键词： graph representation learning graph neural networks graph contrastive learning graph augmentations graph pooling

来源：评论

学校读者我要写书评

暂无评论

A method of embedding a high-resolution image into a large field-of-view image

引用

Multimedia Tools and Applications 2025年第14期84卷 13037-13056页

作者： Dong, Yanmei Pei, Mingtao Jia, Yunde Key Laboratory of Intelligent Information Processing and Graph Processing Tus College of Digit Guangxi University of Science and Technology Liuzhou545006 China Beijing Laboratory of Intelligent Information Technology School of Computer Science Beijing Institute of Technology Beijing100081 China

Images captured by wide-angle cameras or fisheye cameras are with large Field-of-View (FOV) but low resolution, while images captured by conventional cameras are with high-resolution but limited FOV. To handle this contradiction, in this paper, we propose to improve the resolution of a large FOV image by embedding a high-resolution image into it. We present an image embedding method based on non-rigid transformations and mesh optimization. Without any radial distortion correction on large FOV images, we utilize non-rigid transformations to model the mathematical transformation relationship between the high-resolution image and the large FOV image. Meanwhile, we exploit mesh optimization to generate evenly distributed control points for maintaining the shape and content of the large FOV image. To verify the effectiveness of the proposed method, we capture some image pairs from a conventional camera and a large FOV camera, and exploit several classic image alignment methods to implement four baseline methods. The experimental results show that our method preserves the shape and content of the original FOV image from 58% to 82%, which significantly outperforms all baseline methods. The results indicate that the proposed method can effectively embed a high-resolution image into a large FOV image as well as preserve the shape and content of the large FOV image. In other words, the resolution of a large FOV image can be improved via high-resolution image embedding. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

EPAFusion: A Novel Fusion Network Based on Enhancement and Progressive Aware for Infrared-Visible Images in Low-Light

引用

IEEE Sensors Journal 2025年第9期25卷 15378-15390页

作者： Qi, Jianhuan Ni, Bo Yu, Qida Ni, Haibin Zhou, Xiaoyan Chang, Jianhua Nanjing University of Information Science and Technology School of Electronics and Information Engineering Nanjing210044 China Nanjing University of Information Science and Technology Jiangsu Provincial Key Laboratory of Meteorological Observation and Information Processing Nanjing210044 China

Most infrared and visible image fusion algorithms demonstrate satisfactory performance under normal lighting conditions, but often perform poorly in low-light environments because the texture details in visible images are barely visible in the dark. To resolve this issue, we have proposed a novel fusion network based on enhancement and progressive aware for infrared-visible images in low-light (EPAFusion). The EPAFusion network efficiently brightens low-light environments while fusing complementary information from the two modalities. Specifically, we design a degraded illumination disentangled network (DIDNet), which is specially used to eliminate the degraded illumination effect in the visible image and enhance the feature information of the source image. Then, a progressive complementary-aware fusion network (PCAFNet), which is used to generate fused images with rich texture details and high contrast by integrating the complementary information of the dual modal through a progressive approach. Extensive comparative experiments have demonstrated that EPAFusion exhibits superior to state-of-the-art fusion algorithms in terms of subjective and objective evaluation metrics. Particularly, low-light enhancement and dual-modal-aware fusion provide more effective information to the fused image. © 2001-2012 IEEE.

关键词： Image fusion

来源：评论

学校读者我要写书评

暂无评论

Attention-assisted dual-branch interactive face super-resolution network

Cognitive Robotics

引用

Cognitive Robotics 2025年 5卷 77-85页

作者： Wan, Xujie Xu, Siyu Gao, Guangwei Institute of Advanced Technology Nanjing University of Posts and Telecommunications Nanjing China Provincial Key Laboratory for Computer Information Processing Technology Soochow University Suzhou China

We propose a deep learning-based Attention-Assisted Dual-Branch Interactive Network (ADBINet) to improve facial super-resolution by addressing key challenges like inadequate feature extraction and poor multi-scale information handling. ADBINet features a multi-scale encoder-decoder architecture that captures and integrates features across scales, enhancing detail and reconstruction quality. The key to our approach is the Transformer and CNN Interaction Module (TCIM), which includes a Dual Attention Collaboration Module (DACM) for improved local and spatial feature extraction. The Channel Attention Guidance Module (CAGM) refines CNN and Transformer fusion, ensuring precise facial detail restoration. Additionally, the Attention Feature Fusion Unit (AFFM) optimizes multi-scale feature integration. Experimental results demonstrate that ADBINet outperforms existing methods in both quantitative and qualitative facial super-resolution metrics. © 2025

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

DECTNet: A detail enhanced CNN-Transformer network for single-image deraining

Cognitive Robotics

引用

Cognitive Robotics 2025年 5卷 48-60页

作者： Wang, Liping Gao, Guangwei Institute of Advanced Technology Nanjing University of Posts and Telecommunications Nanjing China Provincial Key Laboratory for Computer Information Processing Technology Soochow University Suzhou China

Recently, Convolutional Neural Networks (CNN) and Transformers have been widely adopted in image restoration tasks. While CNNs are highly effective at extracting local information, they struggle to capture global context. Conversely, Transformers excel at capturing global information but often face challenges in preserving spatial and structural details. To address these limitations and harness both global and local features for single-image deraining, we propose a novel approach called the Detail Enhanced CNN-Transformer Network (DECTNet). DECTNet integrates two key components: the Enhanced Residual Feature Distillation Block (ERFDB) and the Dual Attention Spatial Transformer Block (DASTB). In the ERFDB, we introduce a mixed attention mechanism, incorporating channel information-enhanced layers within the residual feature distillation structure. This design facilitates a more effective step-by-step extraction of detailed information, enabling the network to restore fine-grained image details progressively. Additionally, in the DASTB, we utilize spatial attention to refine features obtained from multi-head self-attention, while the feed-forward network leverages channel information to enhance detail preservation further. This complementary use of CNNs and Transformers allows DECTNet to balance global context understanding with detailed spatial restoration. Extensive experiments have demonstrated that DECTNet outperforms some state-of-the-art methods on single-image deraining tasks. Furthermore, our model achieves competitive results on three low-light datasets and a single-image desnowing dataset, highlighting its versatility and effectiveness across different image restoration challenges. © 2025

关键词： Restoration

来源：评论

学校读者我要写书评

暂无评论

Structure-to-word dynamic interaction model for abstractive sentence summarization

引用

Neural Computing and Applications 2025年第9期37卷 6567-6581页

作者： Guan, Yong Guo, Shaoru Li, Ru School of Computer & Information Technology Shanxi University Shanxi China Key Laboratory of Computational Intelligence and Chinese Information Processing of Ministry of Education Shanxi University Shanxi China

Abstractive text summarization aims to capture important information from text and integrate contextual information to guide the summary generation. However, effective integration of important and relevant information remains a challenging problem. Existing graph-based methods only consider either word relations or structure information, but neglect the correlation between them. To simultaneously capture the word relations and structure information from sentences, we propose a novel Structure-to-Word dynamic interaction model for Abstractive Sentence Summarization (SWSum). Specifically, we first represent structure and word relation information of sentences by constructing semantic scenario graph and semantic word relation graph based on FrameNet. We subsequently stack multiple graph-based dynamic interaction layers that iteratively enhance their correlation to learn node representations. Finally, a graph fusion module is designed to obtain better overall graph representations, which provide an attention-based context vector for the decoder to generate summary. Experimental results demonstrate our model outperforms existing state-of-the-art methods on two popular benchmark datasets, i.e., Gigaword and DUC 2004. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

HRAM-VITON: High-Resolution Virtual Try-On with Attention Mechanism

引用

computers, Materials & Continua 2025年第2期82卷 2753-2768页

作者： Yue Chen Xiaoman Liang Mugang Lin Fachao Zhang Huihuang Zhao College of Computer Science and Technology Hengyang Normal UniversityHengyang421002China Hunan Provincial Key Laboratory of Intelligent Information Processing and Application Hengyang421002China

The objective of image-based virtual try-on is to seamlessly integrate clothing onto a target image, generating a realistic representation of the character in the specified attire. However, existing virtual try-on methods frequently encounter challenges, including misalignment between the body and clothing, noticeable artifacts, and the loss of intricate garment details. To overcome these challenges, we introduce a two-stage high-resolution virtual try-on framework that integrates an attention mechanism, comprising a garment warping stage and an image generation stage. During the garment warping stage, we incorporate a channel attention mechanism to effectively retain the critical features of the garment, addressing challenges such as the loss of patterns, colors, and other essential details commonly observed in virtual try-on images produced by existing methods. During the image generation stage, with the aim of maximizing the utilization of the information proffered by the input image, the input features undergo double sampling within the normalization procedure, thereby enhancing the detail fidelity and clothing alignment efficacy of the output image. Experimental evaluations conducted on high-resolution datasets validate the effectiveness of the proposed method. Results demonstrate significant improvements in preserving garment details, reducing artifacts, and achieving superior alignment between the clothing and body compared to baseline methods, establishing its advantage in generating realistic and high-quality virtual try-on images.

关键词： Virtual try-on attention mechanism high-resolution image generation

来源：评论

学校读者我要写书评

暂无评论

Facial micro-expression recognition based on dual-stream fusion network

引用

Multimedia Tools and Applications 2025年 1-14页

作者： Sun, Jiacheng Chen, Changhong Jiangsu Key Laboratory of Intelligent Information Processing and Communication Technology Nanjing University of Posts and Telecommunications No.66 Xin Mofan RD Nanjing210003 China

Micro-expressions(MEs) have emerged as a viable strategy for affective estimation due to their high reliability in emotion detection. In recent years, deep learning methods have been successfully applied to the field of micro-expression recognition. However, extracting and learning features from MEs presents challenges due to their brief duration and subtle intensity. To address these challenges, we propose the dual-stream fusion network (DSFNet). Specifically, we design shallow tokens-to-token vision transformers (T2T-ViT) to effectively capture comprehensive spatial position information. We also fine-tuned the number of ViT encoders and heads to enhance overall model performance. Additionally, the proposed multiscale convolution block (MCB) and attention mechanism modules (AMM) facilitate the effective extraction of detailed and valuable multiscale features from MEs. By employing various sizes of convolutional kernels and attention mechanisms, our approach captures higher-level image information, thereby improving MER accuracy. Finally, we integrate the information obtained from both branches. Performance evaluations on three mainstream ME datasets-SMIC, CASME II, and SAMM-demonstrate that the proposed framework significantly outperforms other advanced methods in micro-expression classification. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Deep semantic network for hyperspectral image classification using enhanced bifurcation and light self-Gaussian attention mechanism

引用

Journal of Applied Remote Sensing 2025年第1期19卷

作者： Wan, Xiaoqing Chen, Feng He, Yupeng Hengyang Normal University College of Computer Science and Technology Hengyang China Hunan Provincial Key Laboratory of Intelligent Information Processing and Application Hengyang China

Convolutional neural networks (CNNs) are widely used in hyperspectral image (HSI) classification due to their strong feature extraction capabilities. Nevertheless, CNN-based classification methods face challenges in capturing deep semantic features effectively, and as the number of layers increases, computational expenses escalate significantly. In addition, spatial locations and spectral bands in HSI data have varying discriminative power. To address these challenges, we propose a novel approach, DSNet, which integrates a multiway attention mechanism with a CNN-based deep semantic network to improve HSI classification. First, we extract shallow abstract features using a module that combines 3D and 2D convolution layers, followed by transformation through a Gaussian-weighted feature tokenizer. Second, we introduce an enhanced bifurcation attention mechanism with dynamically adjustable perceptual ranges to capture spatial semantic information from both horizontal and vertical perspectives. Third, a convolutional multilayer perceptron module is employed to extract high-level feature representations in both spatial and spectral dimensions, which are then refined by a light self-Gaussian attention module. Finally, a linear layer is used to determine the sample labels. Experiments conducted on three publicly available hyperspectral datasets illustrate that the proposed DSNet achieves superior performance compared with several contemporary state-of-the-art methods, even when trained with limited samples. © 2025 Society of Photo-Optical Instrumentation Engineers (SPIE).

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

SURE: Mutually Visible Objects and Self-generated Candidate Labels For Relation Extraction 31

SURE: Mutually Visible Objects and Self-generated Candidate ...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Feng, Yuxuan Chen, Qian Wu, Qianyou Guo, Xin Wang, Suge School of Computer and Information Technology Shanxi University China Key Laboratory of Computational Intelligence and Chinese Information Processing China

ISBN: (纸本)9798891761964

Joint relation extraction models effectively mitigate the error propagation problem inherently present in pipeline models. Nevertheless, joint models face challenges including high computational complexity, complex network architectures, difficult parameter tuning, and notably, limited interpretability. In contrast, recent advances in pipeline relation extraction models (PURE, PL-Marker) have attracted considerable attention due to their lightweight design and high extraction accuracy. A key advancement is the introduction of a marker mechanism, which enhances relation extraction (RE) process by highlighting entities. However, these models primarily focus on generating correct labels. In doing so, they neglect the label selection process. Moreover, they fail to adequately capture the intricate interactions between entity pairs. To overcome these limitations, we develop a Candidate Label Markers (CLMs) mechanism that prioritizes strategic label selection over simple label generation. Furthermore, we facilitate interactions among diverse relation pairs, enabling the identification of more intricate relational patterns. Experimental results show that we achieve a new SOTA performance. Specifically, based on the same Named Entity Recognition (NER) results as theirs, we improve the SOTA methods by 2.5%, 1.9%, 1.3% in terms of strict F1 scores on SciERC, ACE05 and ACE04. © 2025 Association for Computational Linguistics.

关键词： Computational linguistics

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：