检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Yao, Liang Liu, Fan Zhang, Chuanyi Ou, Zhiquan Wu, Ting The College of Computer Science and Software Engineering Hohai University Nanjing210098 China

Knowledge distillation (KD) is an effective method for compressing models in object detection tasks. Due to limited computational capability, UAV-based object detection (UAV-OD) widely adopt the KD technique to obtain lightweight detectors. Existing methods often overlook the significant differences in feature space caused by the large gap in scale between the teacher and student models. This limitation hampers the efficiency of knowledge transfer during the distillation process. Furthermore, the complex backgrounds in UAV images make it challenging for the student model to efficiently learn the object features. In this paper, we propose a novel knowledge distillation framework for UAV-OD. Specifically, a progressive distillation approach is designed to alleviate the feature gap between teacher and student models. Then a new feature alignment method is provided to extract object-related features for enhancing student model’s knowledge reception efficiency. Finally, extensive experiments are conducted to validate the effectiveness of our proposed approach. The results demonstrate that our proposed method achieves state-of-the-art (SoTA) performance in two UAV-OD datasets. Copyright © 2024, The Authors. All rights reserved.

关键词： Unmanned aerial vehicles (UAV)

来源：评论

学校读者我要写书评

暂无评论

AmpNorm: An Effective Style Normalization Method for Single Domain Generalization 7

AmpNorm: An Effective Style Normalization Method for Single ...

引用

7th International Conference on Vision, Image and Signal Processing, ICVISP 2023

作者： Hu, Jingyu Zhong, Tao He, Xilin Xie, Weicheng Song, Siyang Shen, Linlin College of Computer Science and Software Engineering Shenzhen University Shenzhen China University of Leicester LeicesterLE1 7RH United Kingdom

ISBN: (纸本)9781839539848

Deep neural networks have demonstrated remarkable efficacy in numerous computer vision tasks. However, due to the training and testing sets of data coming from different domains, the domain gap limits the performances of deep neural networks. To enhance the generalization performance of deep neural networks, we provide a unique amplitude normalization (AmpNorm) approach to decrease the domain gap from the frequency domain perspective. Specifically, ours AmpNorm collects the amplitude spectrum from the source domain during training and then converts the images from the unseen target domain into those with a style similar to the source domain during testing. Our AmpNorm is simple yet effective, as well as plug-and-play, which is readily implemented into the majority of single-domain generalization (SDG) methods. Extensive results on three public benchmarks demonstrate that our AmpNorm can greatly improve these models’ performance on the invisible target domain. © The Institution of engineering & Technology 2023.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Bi-Orthogonal Projection Learning for Dimensionality Reduction 3

Bi-Orthogonal Projection Learning for Dimensionality Reducti...

引用

3rd International Conference on Intelligent Power and Systems, ICIPS 2023

作者： Liang, Yingyi Fan, Junjun Li, Chunsheng Wen, Jiajun Guangdong University of Technology Guangzhou China Shenzhen University College of Computer Science & Software Engineering Shenzhen China Laboratory for Artificial Intelligence in Design Hong Kong

ISBN: (纸本)9798350328691

This paper proposes a novel bi-orthogonal projection learning (BOPL) for dimensionality reduction (DR) methods, which further extends the existing DR to a more flexible, robust, and sparse embedding framework. Unlike conventional methods that use only one kind of projection for learning in DR, the proposed BOPL introduces two kinds of projection, which are orthogonal and provides the true projection the freedom for more accurate data transformation in subspace learning. Inspired by the observation that the two projections share many similar data structures, the projections are expected to preserve the similarity structure of data by using two different reconstruction ways. The proposed method can also handle data corrupted by noises since a sparse item is employed to compensate for the noises during DR. Several novel unsupervised DR methods (i.e., BOPL_PCA, BOPL_NPE, and BOPL_LPP) are derived from the proposed framework. The results of experiments on the natural and synthetic data sets demonstrate that the proposed methods outperform the existing well-established DR methods. ©2023 IEEE.

关键词： Metadata

来源：评论

学校读者我要写书评

暂无评论

A Self-play and Sentiment-Emphasized Comment Integration Framework Based on Deep Q-Learning in a Crowdsourcing Scenario : Extended Abstract 39

A Self-play and Sentiment-Emphasized Comment Integration Fra...

引用

39th IEEE International Conference on Data engineering, ICDE 2023

作者： Rong, Huan Sheng, Victor S. Ma, Tinghuai Zhou, Yang Al-Rodhaan, Mznah Nanjing University of Information Science & Technology School of Artificial Intelligence Nanjing China Texas Tech University Department of Computer Science Lubbock United States Nanjing University of Information Science & Technology School of Computer & Software Nanjing China Auburn University Department of Computer Science & Software Engineering Auburn United States King Saud University College of Computer & Information Sciences Riyadh Saudi Arabia

ISBN: (纸本)9798350322279

Crowdsourcing is a sourcing model where individuals or organizations obtain goods and services from a large, relatively open and often rapidly evolving group of internet users. The most common way that crowdsourcing can facilitate machine learning is to annotate instances with labels [1]. However, the same instance may have inconsistent class labels, in the eyes of various annotators. Therefore, current efforts in crowdsourcing mainly focus on the truth inference or label integration, to remove inconsistent labels or to alleviate biased labeling. In turn, instances with the integrated labels could facilitate the training on machine learning models. The future direction of crowdsourcing is to apply more fine-grained truth inference methods to different application domains [2]. Consequently, we evolve toward another challenging problem of comment integration. That is, how can we integrate or summarize the core opinions of multiple product comments obtained from users, rather than the discrete labels. © 2023 IEEE.

关键词： Crowdsourcing

来源：评论

学校读者我要写书评

暂无评论

Hybrid GrabCut Hidden Markov Model for Segmentation

引用

computers, Materials & Continua 2022年第7期72卷 851-869页

作者： Soobia Saeed Afnizanfaizal Abdullah N.Z.Jhanjhi Mehmood Naqvi Mehedi Masud Mohammed A.AlZain Department of Software Engineering UniversitiTeknologi Malaysia-UTM81310Malaysia School of Computer Science and Engineering SCE Taylor’s UniversitySubang Jaya47500Malaysia Department of Engineering and Technology Mohak CollegeAlbertaCanada Department of Computer Science College of Computers and Information TechnologyTaif UniversityTaif21944Saudi Arabia Department of Information Technology College of Computers and Information TechnologyTaif UniversityTaif21944Saudi Arabia

Diagnosing data or object detection in medical images is one of the important parts of image segmentation especially those data which is less effective to identify inMRI such as low-grade tumors or cerebral spinal fluid(CSF)leaks in the *** aim of the study is to address the problems associated with detecting the low-grade tumor and CSF in brain is difficult in magnetic resonance imaging(MRI)images and another problem also relates to efficiency and less execution time for segmentation of medical *** tumor and CSF segmentation using trained light field database(LFD)datasets of MRI *** research proposed the new framework of the hybrid k-Nearest Neighbors(k-NN)model that is a combination of hybridization of Graph Cut and Support Vector Machine(GCSVM)and Hidden Markov Model of k-Mean Clustering Algorithm(HMMkC).There are four different methods are used in this research namely(1)SVM,(2)GrabCut segmentation,(3)HMM,and(4)k-mean clustering *** this framework,on the one hand,phase one is to perform the classification of SVM and Graph Cut algorithm to create the maximum margin *** research use GrabCut segmentation method which is the application of the graph cut algorithm and extract the data with the help of scaleinvariant features *** the other hand,in phase two,segment the low-grade tumors and CSF using a method adapted for HMkC and extract the information of tumor or CSF fluid by GCHMkC including iterative conditional maximizing mode(ICMM)with identifying the range of *** evaluation is also performing by the comparison of existing techniques in this *** conclusion,our proposed model gives better results than *** proposed model helps to common man and doctor that can identify their condition of brain *** future,this will model will use for other brain related diseases.

关键词： SVM tumor CSF k-NN MRI grabcut segmentation HMM

来源：评论

学校读者我要写书评

暂无评论

FastNER: Speeding up Inferences for Named Entity Recognition Tasks 9th

FastNER: Speeding up Inferences for Named Entity Recogniti...

引用

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

作者： Zhang, Yuming Gao, Xiangxiang Zhu, Wei Wang, Xiaoling College of Computer Science and Software Engineering Shenzhen University Shenzhen China Shanghai Jiaotong University Shanghai China East China Normal University Shanghai China

ISBN: (纸本)9783031466601

BERT and its variants are the most performing models for named entity recognition (NER), a fundamental information extraction task. We must apply inference speedup methods for BERT-based NER models to be deployed in the industrial setting. Early exiting allows the model to use only the shallow layers to process easy samples, thus reducing the average latency. In this work, we introduce FastNER, a novel framework for early exiting with a BERT biaffine NER model, which supports both flat NER tasks and nested NER tasks. First, we introduce a convolutional bypass module to provide suitable features for the current layer’s biaffine prediction head. This way, an intermediate layer can focus more on delivering high-quality semantic representations for the next layer. Second, we introduce a series of early exiting mechanisms for BERT biaffine model, which is the first in the literature. We conduct extensive experiments on 6 benchmark NER datasets, 3 of which are nested NER tasks. The experiments show that: (a) Our proposed convolutional bypass method can significantly improve the overall performances of the multi-exit BERT biaffine NER model. (b) our proposed early exiting mechanisms can effectively speed up the inference of BERT biaffine model. Comprehensive ablation studies are conducted and demonstrate the validity of our design for our FastNER framework. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

WAN-INT: Cost-Effective In-Band Network Telemetry in WAN With A Performance-aware Path Planner 29

WAN-INT: Cost-Effective In-Band Network Telemetry in WAN Wit...

引用

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

作者： Chen, Simian He, Dongbiao Ma, Xiaopeng Ming, Zhongxing Cui, Laizhong Sangfor Technologies Inc China China National Tobacco Corporation Anhui Province China Shenzhen University College of Computer Science and Software Engineering China China

ISBN: (纸本)9798350330717

With the development of cross-datacenter services, accurate and low-cost network performance measurement enables better traffic scheduling. However, the existing network measurement suffers from a lack of telemetry granularities and excessive costs. Besides, the implementation of INT in the WAN remains undefined. In this work, We propose WAN-INT, a two-phase path orchestration algorithm designed to address the challenges posed by limited communication resources and varying link states in WAN scenarios. By generating a telemetry policy based on this algorithm, it can effectively adapt to the requirements of various applications and network states, achieving a balance between telemetry quality and cost limitations. We conduct experiments in a commercial WAN environment. Results show that WAN-INT outperforms existing schemes and reaches a good compromise between the telemetry quality and cost constraints. Compared with the state-of-the-art telemetry system, WAN-INT effectively reduces by at least 43% of telemetry cost while ensuring telemetry quality. © 2023 IEEE.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Dual-branch Feature Interaction Network with Structure Information Learning for Retinopathy of Prematurity Classification

Dual-branch Feature Interaction Network with Structure Infor...

引用

2023 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2023

作者： Lei, Haijun Zhao, Jia Xie, Hai Liu, Yaling Zhang, Guoming Lei, Baiying Shenzhen University College of Computer Science & Software Engineering Shenzhen China Shenzhen University School of Biomedical Engineering Shenzhen China Shenzhen University Shenzhen Eye Hospital Shenzhen China

ISBN: (纸本)9798350337488

Diagnosing retinopathy of prematurity (ROP) is a time-consuming and complex task, even for experienced clinicians, as it is challenging to determine its specific stages accurately. In this study, we propose an advanced dual-branch feature interaction network for predicting the stages of ROP using color fundus photographs. Specifically, the proposed network includes a Vision Transformer (ViT) branch and a convolutional neural network (CNN) branch, which are used to capture global contextual information and express local detail features, respectively. To improve the efficiency of ViT, we introduce a cascaded group attention (CGA) module feeding attention heads with different splits of the full feature, which not only saves computation cost but also improves attention diversity. The semantic features extracted from both the Transformer and CNN branches are fused through the branch feature interaction (BFI) module, allowing us to leverage the unique characteristics of both branches to optimize ROP feature representations comprehensively. Afterwards, we further design a Transformer block with structure information learning (SIL) to gather ROP-related semantic information from high-level features, gradually constructing ROP feature information structure to highlight important regions and improve the model's discriminative ability for different lesion feature structures. Our extensive experiments on both clinical and public datasets produce promising results, showcasing the outstanding performance of our method. © 2023 IEEE.

关键词： dual-branch feature interaction ROP diagnosis structure information learning

来源：评论

学校读者我要写书评

暂无评论

Neighbor Discovery in a Multi-Transceiver Free-Space-Optical Ad Hoc Network 19

Neighbor Discovery in a Multi-Transceiver Free-Space-Optical...

引用

19th IEEE Annual Consumer Communications and Networking Conference, CCNC 2022

作者： Vazquez-Estrada, Jessica Bhunia, Suman Khan, Mahmudur Qian, Yicheng Huu, Nero Tran Miami University Department of Computer Science and Software Engineering OxfordOH United States York College of Pennsylvania Electrical and Computer Engineering and Computer Science Department YorkPA United States

ISBN: (纸本)9781665431620

In this paper, we present a novel neighbor discovery method for a wireless ad hoc network where each node is equipped with a Free-Space-Optical (FSO) transceivers capable of electronic beam switching. Directional neighbor discovery can be very challenging due to the requirement of strict line-of-sight (LOS) alignment and can result in long and impractical discovery time if the nodes in the network do not have prior information about each other's location. Our method guarantees that a node can discover all of its neighbors within one 360 FSO beam sweep. We present a preliminary prototype of a FSO transceiver module capable of electronic beam steering. Through experiments using the developed prototype, we demonstrate that the proposed method helps in discovering neighbor nodes with minimal delay. © 2022 IEEE.

关键词： Optical communication

来源：评论

学校读者我要写书评

暂无评论

DD-UNet: Densely Dilated U-Net for Curvilinear Structure Segmentation in Fundus Image

DD-UNet: Densely Dilated U-Net for Curvilinear Structure Seg...

引用

2023 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2023

作者： Zhang, Yindong Chen, Jie Wang, Li Chen, Miaohong Zhang, Guoming Li, Jianqiang Shenzhen University College of Computer Science and Software Engineering China Shenzhen University National Engineering Laboratory for Big Data System Computing Technology China Shenzhen Eye Hospital China

ISBN: (纸本)9798350337488

Retinopathy of Prematurity (ROP) is a retina disorder that affects premature infants with lower weights. If the patient cannot get the treatment in time when the illness reaches the last stage, irreversible vision loss will be caused. Nevertheless, there has been relatively little consideration given to the segmentation of the ridge, the key clinical characteristic of the illness. Additionally, existing research has not adequately addressed several segmentation issues, such as fragmentary topology, class imbalance, and false positives. This paper proposes a Densely Dilated U-Net (DD-UNet) improved from U-Net to tackle these challenges. Furthermore, the post-processing techniques based on the spatial relationship between vessels and ridges, along with the relative pixel counts of ridges and false positive results is integrated to mitigate false positive results in the predicted ridge. To enhance the precision of thin vessel, a sliding window sampling method is introduced for refined training. Compared with the state-of-the-art models in medical image segmentation, DD-UNet performs well in curvilinear structure segmentation of fundus image. For instance, our DD-UNet outperforms the Attention U-Net by 6.26% in terms of sensitivity and exhibits a 1.85% higher dice score in ridge segmentation. © 2023 IEEE.

关键词： curvilinear structure segmentation fundus image Retinopathy of Prematurity (ROP) ridge segmentation vessel segmentation

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：