检索结果-南通市图书馆

维普期刊数据库

同方期刊数据库

CPT: a pre-trained unbalanced transformer for both Chinese language understanding and generation

在线全文

学校读者我要写书评

暂无评论

Science China(information Sciences) 2024年第5期67卷 43-55页

作者： Yunfan SHAO Zhichao GENG Yitao LIU Junqi DAI Hang YAN Fei YANG Zhe LI Hujun BAO Xipeng QIU School of Computer Science Fudan University Shanghai key laboratory of Intelligent information processing Fudan University Zhejiang Lab

In this paper, we take the advantage of previous pre-trained models(PTMs) and propose a novel Chinese pre-trained unbalanced transformer(CPT). Different from previous Chinese PTMs, CPT is designed to utilize the shared knowledge between natural language understanding(NLU) and natural language generation(NLG) to boost the performance. CPT consists of three parts: a shared encoder, an understanding decoder, and a generation decoder. Two specific decoders with a shared encoder are pretrained with masked language modeling(MLM) and denoising auto-encoding(DAE) tasks, *** the partially shared architecture and multi-task pre-training, CPT can(1) learn specific knowledge of both NLU or NLG tasks with two decoders and(2) be fine-tuned flexibly that fully exploits the potential of the model. Moreover, the unbalanced transformer saves the computational and storage cost, which makes CPT competitive and greatly accelerates the inference of text generation. Experimental results on a wide range of Chinese NLU and NLG tasks show the effectiveness of CPT.

关键词： pre-trained model transformer language model generation unified model

Lightweight Object Detection Networks for UAV Aerial Images Based on YOLO

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2024年第04期33卷 997-1009页

作者： Yanshan LI Jiarong WANG Kunhua ZHANG Jiawei YI Miaomiao WEI Lirong ZHENG Weixin XIE ATR National key laboratory of Defense Technology Shenzhen University Guangdong key laboratory of Intelligent information processing Shenzhen University

Existing high-precision object detection algorithms for UAV(unmanned aerial vehicle) aerial images often have a large number of parameters and heavy weight, which makes it difficult to be applied to mobile *** propose three YOLO-based lightweight object detection networks for UAVs, named YOLO-L, YOLO-S, and YOLO-M, respectively. In YOLO-L, we adopt a deconvolution approach to explore suitable upsampling rules during training to improve the detection accuracy. The convolution-batch normalization-Si LU activation function(CBS)structure is replaced with Ghost CBS to reduce the number of parameters and weight, meanwhile Maxpool maximum pooling operation is proposed to replace the CBS structure to avoid generating parameters and weight. YOLO-S greatly reduces the weight of the network by directly introducing CSPGhost Neck residual structures, so that the parameters and weight are respectively decreased by about 15% at the expense of 2.4% mAP. And YOLO-M adopts the CSPGhost Neck residual structure and deconvolution to reduce parameters by 5.6% and weight by 5.7%, while m AP only by 1.8%. The results show that the three lightweight detection networks proposed in this paper have good performance in UAV aerial image object detection task.

关键词： Training YOLO Deconvolution Accuracy Object detection Autonomous aerial vehicles Feature extraction

同方期刊数据库

A Comprehensive Overview of CFN From a Commonsense Perspective

在线全文

学校读者我要写书评

暂无评论

Machine Intelligence Research 2024年第2期21卷 239-256页

作者： Ru Li Yunxiao Zhao Zhiqiang Wang Xuefeng Su Shaoru Guo Yong Guan Xiaoqi Han Hongyan Zhao key laboratory of Computational Intelligence and Chinese information processing of Ministry of Education School of Computer and Information TechnologyShanxi UniversityTaiyuan 030006China

Chinese FrameNet(CFN)is a scenario commonsense knowledge base(CKB)that plays an important role in research on Chinese language *** is based on the theory of frame semantics and English FrameNet(FN).The CFN knowledge base contains a wealth of scenario commonsense knowledge,including frames,frame elements,and frame relations,as well as annotated instances with rich scenario-related labels on Chinese sentences and *** this paper,we conduct a comprehensive overview of CFN from a commonsense perspective,covering topics such as scenario commonsense representation,CFN resources,and its *** also summarize recent breakthroughs and identify future research ***,we introduce the concept of scenario commonsense,including its definitions,examples,and representation methods,with a focus on the relationship between scenario commonsense and the frame concept in *** addition,we provide a comprehensive overview of CFN resources and their applications,highlighting the newly proposed frame-based discourse representation and a human-machine collaboration framework for expanding the CFN ***,we explore emerging topics such as expanding the CFN resource,improving the interpretability of machine reading comprehension,and using scenario CKBs for text generation.

关键词： Chinese FrameNet(CFN) commonsense scenario commonsense frame knowledge

Track-before-detect for bistatic radar based on pseudo-spectrum

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Aeronautics 2024年第1期37卷 236-246页

作者： Tao HAN Gongjian ZHOU School of Electronics and information Engineering Harbin Institute of TechnologyHarbin 150001China key laboratory of Marine Environmental Monitoring and information processing Ministry of Industry and Information TechnologyHarbin 150001China

Track-Before-Detect(TBD) is an efficient method to detect dim targets for radars. Conventional TBD usually follows an approximate motion model of the target, which may cause an inaccurate integration of the target energy. A TBD technique on basis of pseudo-spectrum in mixed coordinates adopting an accurate motion model for bistatic radar system is developed in this *** predicted position in bistatic polar plane is derived according to a nonlinear function that exactly describes the constant Cartesian velocity motion. Then around the predicted position, a pseudo-spectrum is formulated and its samples are accumulated to the integration frame for energy integration. The evolution of the state and the procedure of accumulation of the target energy are derived elaborately. The superior performance of the proposed method is demonstrated by some simulations.

关键词： Weak target detection Bistatic radar Track-before-detect Pseudo-spectrum Multiframe integration

维普期刊数据库

On Isodual Double Toeplitz Codes

在线全文

学校读者我要写书评

暂无评论

Journal of Systems Science & Complexity 2024年第5期37卷 2196-2206页

作者： SHI Minjia XU Li SOLé Patrick key laboratory of Intelligent Computing and Signal processing Ministry of Education School of Mathematical Sciences Anhui University State key laboratory of information Security Institute of Information Engineering Chinese Academy of Sciences CNRS University of Aix Marseille Centrale Marseille

Double Toeplitz（shortly DT） codes are introduced here as a generalization of double circulant codes. The authors show that such a code is isodual, hence formally self-dual（FSD）. FSD codes form a far-reaching generalization of self-dual codes, the most important class of codes of rate one-half. Self-dual DT codes are characterized as double circulant or double negacirculant. Likewise, even binary DT codes are characterized as double circulant. Numerical examples obtained by exhaustive search show that the codes constructed have best-known minimum distance, up to one unit, amongst formally self-dual codes, and sometimes improve on the known values. For q = 2, the authors find four improvements on the best-known values of the minimum distance of FSD codes. Over F4an explicit construction of DT codes, based on quadratic residues in a prime field, performs equally well. The authors show that DT codes are asymptotically good over Fq. Specifically, the authors construct DT codes arbitrarily close to the asymptotic Varshamov-Gilbert bound for codes of rate one half.

关键词： Double circulant codes double Toeplitz codes formally self-dual codes isodual codes

Self-attention Guidance Based Crowd Localization and Counting

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Machine Intelligence Research 2024年第5期21卷 966-982页

作者： Zhouzhou Ma Guanghua Gu Wenrui Zhao School of information Science and Engineering Yanshan UniversityQinhuangdao066000China Hebei key laboratory of information Transmission and Signal processing Qinhuangdao066000China

Most existing studies on crowd analysis are limited to the level of counting,which cannot provide the exact location of *** paper proposes a self-attention guidance based crowd localization and counting network(SA-CLCN),which can simultaneously locate and count *** take the form of object detection,using the original point annotations of crowd datasets as supervision to train the ***,the center point coordinate of each head as well as the number of crowds are ***,to cope with the spatial and positional variations of the crowd,the proposed method introduces transformer to construct a globallocal feature extractor(GLFE)together with the convolutional *** establishes the near-to-far dependency between elements so that the global context and local detail features of the crowd image can be extracted ***,this paper designs a pyramid feature fusion module(PFFM)to fuse the global and local information from high level to low level to obtain a multiscale feature *** downstream tasks,this paper predicts candidate point offsets and confidence scores by a simple regression header and classification *** addition,the Hungarian algorithm is used to match the predicted point set and the labelled point set to facilitate the calculation of *** proposed network avoids the errors or higher costs associated with using traditional density maps or bounding box ***,we have conducted extensive experiments on several crowd datasets,and the proposed method has produced competitive results in both counting and localization.

关键词： Crowd localization crowd counting transformer point supervision object detection

Co-salient object detection with iterative purification and predictive optimization

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

虚拟现实与智能硬件 2024年第5期6卷 396-407页

作者： Yang WEN Yuhuan WANG Hao WANG Wuzhen SHI Wenming CAO Guangdong Provincial key laboratory of Intelligent information processing College of Electronics and Information Engineering Shenzhen University Shenzhen China 518060

BackgroundCo-salient object detection (Co-SOD) aims to identify and segment commonly salient objects in a set of related images. However, most current Co-SOD methods encounter issues with the inclusion of irrelevant information in the co-representation. These issues hamper their ability to locate co-salient objects and significantly restrict the accuracy of detection. MethodsTo address this issue, this study introduces a novel Co-SOD method with iterative purification and predictive optimization (IPPO) comprising a common salient purification module (CSPM), predictive optimizing module (POM), and diminishing mixed enhancement block (DMEB). ResultsThese components are designed to explore noise-free joint representations, assist the model in enhancing the quality of the final prediction results, and significantly improve the performance of the Co-SOD algorithm. Furthermore, through a comprehensive evaluation of IPPO and state-of-the-art algorithms focusing on the roles of CSPM, POM, and DMEB, our experiments confirmed that these components are pivotal in enhancing the performance of the model, substantiating the significant advancements of our method over existing benchmarks. Experiments on several challenging benchmark co-saliency datasets demonstrate that the proposed IPPO achieves state-of-the-art performance.

关键词： Co-salient object detection Saliency detection Iterative method Predictive optimization

A review of medical ocular image segmentation

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

虚拟现实与智能硬件（中英文） 2024年第3期6卷 181-202页

作者： Lai WEI Menghan HU Shanghai key laboratory of Multidimensional information processing East China Normal UniversityShanghai 200241China

Deep learning has been extensively applied to medical image segmentation,resulting in significant advancements in the field of deep neural networks for medical image segmentation since the notable success of U Net in ***,the application of deep learning models to ocular medical image segmentation poses unique challenges,especially compared to other body parts,due to the complexity,small size,and blurriness of such images,coupled with the scarcity of *** article aims to provide a comprehensive review of medical image segmentation from two perspectives:the development of deep network structures and the application of segmentation in ocular ***,the article introduces an overview of medical imaging,data processing,and performance evaluation ***,it analyzes recent developments in U-Net-based network ***,for the segmentation of ocular medical images,the application of deep learning is reviewed and categorized by the type of ocular tissue.

关键词： Medical image segmentation Orbit Tumor U-Net Transformer

维普期刊数据库

A Tabletop Nano-CT Image Noise Reduction Network Based on 3-Dimensional Axial Attention Mechanism

在线全文

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2024年第7期80卷 1711-1725页

作者： Huijuan Fu Linlin Zhu ChunhuiWang Xiaoqi Xi Yu Han Lei Li Yanmin Sun Bin Yan Henan key laboratory of Imaging and Intelligent processing PLA Strategic Support Force Information Engineering UniversityZhengzhou450000China

Nano-computed tomography(Nano-CT)is an emerging,high-resolution imaging ***,due to their low-light properties,tabletop Nano-CT has to be scanned under long exposure conditions,which the scanning process is *** 3D reconstruction data,this paper proposed a lightweight 3D noise reduction method for desktop-level Nano-CT called AAD-ResNet(Axial Attention DeNoise ResNet).The network is framed by theU-net *** encoder and decoder are incorporated with the proposed 3D axial attention mechanism and residual dense *** layer of the residual dense block can directly access the features of the previous layer,which reduces the redundancy of parameters and improves the efficiency of network *** 3D axial attention mechanism enhances the correlation between 3D information in the training process and captures the long-distance *** can improve the noise reduction effect and avoid the loss of image structure *** results show that the network can effectively improve the image quality of a 0.1-s exposure scan to a level close to a 3-s exposure,significantly shortening the sample scanning time.

关键词： Deep learning tabletop Nano-CT image denoising 3D axial attention mechanism