检索结果-内蒙古大学图书馆

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Ying Zhang Shuai Guo Dantong Shi Jinhai Xiang College of Informatics Huazhong Agricultural University Wuhan China Agricultural Bioinformatics Key Laboratory of Hubei Province Key Laboratory of Smart Farming for Agricultural Animals Ministry of Agriculture Hubei Engineering Technology Research Center of Agricultural Big Data Engineering Research Center of Intelligent Technology for Agriculture Ministry of Education Huazhong Agricultural University Wuhan China

ISBN: (数字)9798350386226

ISBN: (纸本)9798350386233

The core of the image registration task is to accurately extract and compare the spatial feature information between moving and fixed images. The model must not only be able to capture features inside an image, but also be able to distinguish similarities and differences between moving image and fixed image. Although UNet and its variant networks are widely used in the field of registration, they still present challenges in terms of fine-grained and multi-level feature extraction. In order to address these issues, we propose a dual-decoder UNet network based on contrastive learning. For enhancing the discriminant ability of the network at the feature level, the proposed model applies contrastive learning to the encoder stage of UNet, which strengthens the differential learning in the process of feature extraction. In addition, a multi-level feature fusion method is introduced in the encoder to consider the feature differences at each level more comprehensively. Experimental results show that the proposed method has achieved a significant improvement in the registration accuracy. On the LPBA40 dataset, compared with Affine, VoxelMorph and CLMorph, the Dice scores were increased by 15.2%, 2.5% and 1.9%, respectively, which verified the effectiveness of the proposed method. Our code implementation is publicly available at https://***/LaneAzur/DD-CLUNet.

关键词： Learning systems Image registration Accuracy Deformation Contrastive learning Feature extraction Transformers Decoding data mining Biomedical imaging

来源：评论

学校读者我要写书评

暂无评论

LDVNet: Lightweight and Detail-Aware Vision Network for Image Recognition Tasks in Resource-Constrained Environments 29

LDVNet: Lightweight and Detail-Aware Vision Network for Imag...

引用

29th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2023

作者： Liu, Yuchen Liang, Hu Qiu, Ran Zhao, Shengrong Qilu University of Technology Shandong Academy of Sciences Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center Jinan China Qilu University of Technology Shandong Academy of Sciences Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and Technology Jinan China Shandong Fundamental Research Center for Computer Science Shandong Provincial Key Laboratory of Computer Networks Jinan China

ISBN: (纸本)9798350330717

In many underwater application scenarios, recognition tasks need to be executed promptly on computationally limited platforms. However, models designed for this field often exhibit spatial locality, and existing works lack the ability to capture crucial details in images. Therefore, a lightweight and detail-aware vision network (LDVNet) for resource-constrained environments is proposed to overcome the limitations of these approaches. Firstly, in order to enhance the accuracy of target image recognition, we introduce transformer modules to acquire global information, thus addressing the issue of spatial locality inherent in traditional convolutional neural networks (CNNs). Secondly, to maintain the network's lightweight nature, we integrate the transformer module with convolutional operations, thereby mitigating the substantial parameter and floating point operations (FLOPs) overhead. Thirdly, for the efficient extraction of crucial fine-grained details from feature maps, we have devised a channel and spatial attention module (C&SA). This module aids in recognizing intricate and fine-grained visual tasks and enhances image understanding. It is seamlessly integrated into LDVNet with nearly negligible parameter overhead. The experimental results demonstrate that LDVNet outperforms other lightweight networks and hybrid networks in different recognition tasks, while being suitable for resource-constrained environments. © 2023 IEEE.

关键词： Detail-aware Image recognition Lightweight

来源：评论

学校读者我要写书评

暂无评论

Secrecy Performance Analysis of Artificial Noise Assisted Secure Transmission for MISO systems with Multiple Eavesdroppers

Secrecy Performance Analysis of Artificial Noise Assisted Se...

引用

IEEE International Conference on Signal Processing, Communications and Computing (ICSPCC)

作者： Panpan Xu Weijie Tan Kun Niu Zhen Li Changgen Peng School of Mathematics and Statistics Guizhou University Guiyang China Key Laboratory of Advanced Manufacturing Technology of Ministry of Education Guizhou China Guizhou Big Data Academy Guizhou University Guiyang China State Key Laboratory of Public Big Data College of Computer Science and Technology Guizhou University Guiyang China College of Big Data and Information Engineering Guizhou University Guiyang China

In the paper, we investigate the secure communication of multiple-input single-output (MISO) systems with multiple eavesdroppers. We jointly design the beamforming (BF) and the artificial noise (AN) in MISO systems with multiple eavesdroppers to enhance the secrecy rate. Furthermore, under the secrecy interrupt constraint, we obtain the calculation formulas of the secrecy interrupt probability and the optimal secrecy rate. The simulation results demonstrate that AN with a higher dimension can effectively obstruct eavesdroppers and enhance the system's optimal secrecy rate. In addition, it is proved that the 1/4 full-dimension AN's secrecy rate close to that of the full-dimension AN in the system with multiple eavesdroppers, allowing the trade-off between system secrecy performance and artificial noise design complexity to be realized.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Learning Relation-Specific Representations for Few-shot knowledge Graph Completion

arXiv

引用

arXiv 2022年

作者： Li, Yuling Yu, Kui Zhang, Yuhong Wu, Xindong Key Laboratory of Knowledge Engineering with Big Data of Ministry of Education Hefei University of Technology Hefei230601 China School of Computer Science and Information Engineering Hefei University of Technology Hefei230601 China

Recent years have witnessed increasing interest in few-shot knowledge graph completion (FKGC), which aims to infer unseen query triples for a few-shot relation using a few reference triples about the relation. The primary focus of existing FKGC methods lies in learning relation representations that can reflect the common information shared by the query and reference triples. To this end, these methods learn entity-pair representations from the direct neighbors of head and tail entities, and then aggregate the representations of reference entity pairs. However, the entity-pair representations learned only from direct neighbors may have low expressiveness when the involved entities have sparse direct neighbors or share a common local neighborhood with other entities. Moreover, merely modeling the semantic information of head and tail entities is insufficient to accurately infer their relational information especially when they have multiple relations. To address these issues, we propose a Relation-Specific Context Learning (RSCL) framework, which exploits graph contexts of triples to learn global and local relation-specific representations for few-shot relations. Specifically, we first extract graph contexts for each triple, which can provide long-term entity-relation dependencies. To encode the extracted graph contexts, we then present a hierarchical attention network to capture contextualized information of triples and highlight valuable local neighborhood information of entities. Finally, we design a hybrid attention aggregator to evaluate the likelihood of the query triples at the global and local levels. Experimental results on two public datasets demonstrate that RSCL outperforms state-of-the-art FKGC methods. © 2022, CC BY.

关键词： knowledge graph

来源：评论

学校读者我要写书评

暂无评论

GAF-GAN: Gated Attention Feature Fusion Image Inpainting Network Based on Generative Adversarial Network 29

GAF-GAN: Gated Attention Feature Fusion Image Inpainting Net...

引用

29th IEEE International Conference on Parallel and Distributed Systems, ICPADS 2023

作者： Qiu, Ran Liang, Hu Liu, Yuchen Zhao, Shengrong Qilu University of Technology Shandong Academy of Sciences Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center Jinan China Qilu University of Technology Shandong Academy of Sciences Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and Technology Jinan China Shandong Fundamental Research Center for Computer Science Shandong Provincial Key Laboratory of Computer Networks Jinan China

ISBN: (纸本)9798350330717

Image inpainting, which aims to reconstruct reasonably clear and realistic images from known pixel information, is one of the core problems in computer vision. However, due to the complexity and variability of the underwater environment, the inability to extract valid pixel points and insufficient correlation between feature information in existing image inpainting techniques lead to blurring in the generated images. Therefore, a novel gated attention feature fusion image inpainting network based on generative adversarial networks (GAF-GAN) is proposed. The accuracy of feature similarity matching depends heavily on the validity of the information contained in the features. On the one hand, gating values are dynamically generated by gated convolution to reduce the interference of invalid information. On the other hand, semantic information at distant locations in an image is accurately acquired by the attention mechanism. For these reasons, we designed an improved gated attention mechanism. Gated attention mechanism make the network focus on effective information such as high-frequency texture and color fidelity of restored images. In addition, the dense feature fusion module is added to expand the overall receptive field of the network to fully learn the image features. Experimental results show that the proposed method can effectively repair defective images with complex texture structures and improve the reality and integrity of image details and structures. © 2023 IEEE.

关键词： Dense feature fusion module Gated attention mechanism Image inpainting Underwater environment

来源：评论

学校读者我要写书评

暂无评论

Semi-Supervised Medical Image Segmentation Based on Frequency Domain Aware Stable Consistency Regularization

引用

Journal of imaging informatics in medicine 2025年 2025 Jan 22页

作者： Yihao Ouyang Peipei Li Haixiang Zhang Xuegang Hu Key Laboratory of Knowledge Engineering with Big Data (the Ministry of Education of China) Hefei University of Technology Hefei 230009 Anhui China. School of Computer Science and Information Engineering Hefei University of Technology Hefei 230009 Anhui China. School of Computer Science and Information Engineering Hefei University of Technology Hefei 230009 Anhui China. peipeili@***. Center for Big Data and Population Health of IHM Anhui Medical University Hefei Anhui China. peipeili@***. Center for Big Data and Population Health of IHM Anhui Medical University Hefei Anhui China. Computer Centre The Second People's Hospital of Hefei Hefei 230011 Anhui China. Anhui Province Key Laboratory of Industry Safety and Emergency Technology Hefei University of Technology Hefei 230009 Anhui China.

With the advancement of deep learning models nowadays, they have successfully applied in the semi-supervised medical image segmentation where there are few annotated medical images and a large number of unlabeled ones. A representative approach in this regard is the semi-supervised method based on consistency regularization, which improves model training by imposing consistency constraints (perturbations) on unlabeled data. However, the perturbations in this kind of methods are often artificially designed, which may introduce biases unfavorable to the model learning in the handling of medical image segmentation. On the other hand, the majority of such methods often overlook the supervision in the Encoder stage of training and primarily focus on the outcomes in the later stages, potentially leading to chaotic learning in the initial phase and subsequently impacting the learning process of the model in the later stages. At the meanwhile, they miss the intrinsic spatial-frequency information of the images. Therefore, in this study, we propose a new semi-supervised medical image segmentation approach based on frequency domain aware stable consistency regularization. Specifically, to avoid the bias introduced by artificially setting perturbations, we first utilize the inherent frequency domain information of images, including both high and low frequencies, as the consistency constraint. Secondly, we incorporate supervision in the Encoder stage of model training to ensure that the model does not fail to learn due to the disruption of the original feature space caused by strong augmentation. Finally, extensive experimentation validates the effectiveness of our semi-supervised approach.

关键词： Consistency regularization Frequency domain Medical image segmentation Semi-supervision

来源：评论

学校读者我要写书评

暂无评论

EvoSampling: A Granular Ball-based Evolutionary Hybrid Sampling with knowledge Transfer for Imbalanced Learning

arXiv

引用

arXiv 2024年

作者： Pei, Wenbin Dai, Ruohao Xue, Bing Zhang, Mengjie Zhang, Qiang Cheung, Yiu-Ming Xia, Shuyin The School of Computer Science and Technology Dalian University of Technology Dalian116024 China Ministry of Education Dalian116024 China The School of Engineering and Computer Science Victoria University of Wellington PO Box 600 Wellington6140 New Zealand The Department of Computer Science Hong Kong Baptist University Hong Kong The Chongqing Key Laboratory of Computational Intelligence Key Laboratory of Big Data Intelligent Computing Key Laboratory of Cyberspace Big Data Intelligent Security Ministry of Education Chongqing University of Posts and Telecommunications Chongqing400065 China

Class imbalance would lead to biased classifiers that favor the majority class and disadvantage the minority class. Unfortunately, from a practical perspective, the minority class is of importance in many real-life applications. Hybrid sampling methods address this by oversampling the minority class to increase the number of its instances, followed by undersampling to remove low-quality instances. However, most existing sampling methods face difficulties in generating diverse high-quality instances and often fail to remove noise or low-quality instances on a larger scale effectively. This paper therefore proposes an evolutionary multi-granularity hybrid sampling method, called EvoSampling. During the oversampling process, genetic programming (GP) is used with multi-task learning to effectively and efficiently generate diverse high-quality instances. During the undersampling process, we develop a granular ball-based undersampling method that removes noise in a multi-granular fashion, thereby enhancing data quality. Experiments on 20 imbalanced datasets demonstrate that EvoSampling effectively enhances the performance of various classification algorithms by providing better datasets than existing sampling methods. Besides, ablation studies further indicate that allowing knowledge transfer accelerates the GP's evolutionary learning process. Copyright © 2024, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A calibration method of structural model for multi-FOV star tracker based on theodolite crosshair imaging

A calibration method of structural model for multi-FOV star ...

引用

2021 International Conference on Optical Instruments and Technology: Optoelectronic Measurement Technology and Systems

作者： Tian, Lingfeng Jiang, Jie Yang, Jisan School of Instrumentation and Optoelectronic Engineering Beihang University No. 37 Xueyuan Road Beijing100191 China Precision Opto-mechatronics Technology Key Laboratory of Education Ministry Beijing China Beijing Advanced Innovation Center for Big Data-Based Precision Medicine Beijing China

ISBN: (纸本)9781510655690

Compared with the traditional single field of view (FOV) star tracker, the multi-FOV star tracker has the advantages of equal measurement accuracy on three axes and better dynamic performance. The measurement accuracy of a multi-FOV star tracker is directly determined by the accuracy of the structural model. However, existing structural model calibration methods cannot be applied to the high precision multi-FOV star trackers with large size and weight. To solve this issue, a calibration method of the structural model for the multi-FOV star tracker based on theodolite crosshair imaging is proposed. An imaging model of theodolite crosshair in star tracker is established which elaborates the relationship between the theodolite angles and the star tracker images. Multiple theodolites are utilized to pair each FOV of the star tracker. In each pair, through collecting the angles measured by the theodolite and the images captured by the star tracker, the rotation between the theodolite frame and the star tracker FOV frame could be solved. Additionally, the rotations between theodolite frames are obtained by mutual collimation of theodolites. Finally, the structural model which contains the rotations between the different FOV frames is acquired by merging the above rotations. Structural model calibration experiments of a multi-FOV star tracker with a large size and weight have been conducted. The experimental result indicated that the mean star angular distance error between FOVs was less than 10 arcseconds. The accuracy of the calibration result met the practical requirements. The proposed method is free from the in uence of the size and weight of multi-FOV star trackers and maintains high calibration accuracy. © 2022 SPIE.

关键词： Stars

来源：评论

学校读者我要写书评

暂无评论

Automatic Detection of Active Geohazards with Millimeter-to-Meter-Scale Deformation and Quantitative Analysis of Factors Influencing Spatial Distribution: A Case Study in the Hexi Corridor, China

SSRN

引用

SSRN 2024年

作者： Chen, Bo Li, Zhenhong Song, Chuang Yu, Chen Zhu, Wu Liu, Zhenjiang Han, Bingquan Du, Jiantao Zhang, Chenglong Xu, Fu Peng, Jianbing College of Geological Engineering and Geomatics Chang’an University Xi’an710054 China Key Laboratory of Loess Xi’an710054 China Big Data Center for Geosciences and Satellites Xi’an710054 China Key Laboratory of Western China’s Mineral Resources and Geological Engineering Ministry of Education Xi’an710054 China Key Laboratory of Ecological Geology and Disaster Prevention Ministry of Natural Resources Xi’an710054 China

With the escalation of global climate change and human activity, geohazards become increasingly frequent which cause severe casualties and property losses to local communities. To alleviate this situation and provide scientific guidance for risk reduction, it is imperative to address some of the basic questions related to geohazards, including: i) how to detect active geohazards ( AGs) rapidly and automatically over a wide area;ii) how to determine the region with a high level of hazard activity;iii) what are the primary conditioning factors (CFs) of AGs;and iv) do factors operate independently or are they interconnected. To tackle these issues, we propose a universally applicable framework for wide area automated detection of AGs. The framework is based on multi-source Earth observations which capture surface deformation ranging from millimeters to meters. Our study has focused on the Hexi Corridor (HXC) in Gansu Province, China, covering an area of 210,000 km² with a length of 1,100 km. First, we construct an AGs database for the HXC with high automatic and rapid update capabilities, including a total of 4,492 AGs (3,652 active landslides and 840 land subsidence areas). Second, using the Geographic Detectors method, we determine the primary CFs including elevation, land surface temperature, and precipitation. We find that faults exert greater control over very slow-moving landslides, but are less effective over slow-moving landslides. Third, we analyzed the interactive effects of dual CFs on geohazard actives. Any interaction effect of dual CFs contributes to the bivariate enhancement of geohazard activity. This study significantly enhances the capabilities of the wide area automated detection of AGs, and provides a crucial dataset for hazard prediction and mitigation along the HXC. © 2024, The Authors. All rights reserved.

关键词： Hazards

来源：评论

学校读者我要写书评

暂无评论

Structure optimization for airborne multi-FOV star trackers

Structure optimization for airborne multi-FOV star trackers

引用

2021 International Conference on Optical Instruments and Technology: Optoelectronic Measurement Technology and Systems

作者： Yang, Jisan Jiang, Jie Tian, Lingfeng School of Instrumentation and Optoelectronic Engineering Beihang University No. 37 Xueyuan Road Beijing100191 China Precision Opto-mechatronics Technology Key Laboratory of Education Ministry Beijing China Beijing Advanced Innovation Center for Big Data-Based Precision Medicine Beijing China

ISBN: (纸本)9781510655690

An airborne multiple-field-of-view (multi-FOV) star tracker operating inside the atmosphere has particular limitations in observation. It is necessary to optimize its structure parameters for improving its attitude determination and reliability in the working circumstances. In this paper, performance simulations for different multi-FOV structures are carried out. In the simulation design, the terrain occlusion, stellar atmospheric refraction, and atmospheric extinction are the main considerations. When conducting the simulation experiments, within the permitted attitude ranges, abundant random attitude of a star tracker was generated for performance testing, including three-axis attitude error and stellar detection probability. The results show that, with no refraction exists, when the tilt angle of the boresight of each FOV is at 40°~45°, regardless of which structural layout is adopted, both the attitude measurement accuracy and the stellar detect probability of single FOV is relatively high. With refraction exists, the tilt angles of the boresights are larger, the attitude measurement error is greater. For an airborne multi-FOV star tracker, a FOV with 0° tilt angle is necessary to promote its reliability. © 2022 SPIE.

关键词： Star trackers

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：