检索结果-内蒙古大学图书馆

Enhanced Multi-Scale Object Detection Algorithm for Foggy Traffic Scenarios

computers, Materials & Continua 2025年第2期82卷 2451-2474页

作者： Honglin Wang Zitong Shi Cheng Zhu School of Artificial Intelligence Nanjing University of Information Science and TechnologyNanjing210044China School of Computer Science Nanjing University of Information Science and TechnologyNanjing210044China School of Electrical&Computer Engineering University of Illinois at Urbana ChampaignUrbanaIL 61801USA

In foggy traffic scenarios, existing object detection algorithms face challenges such as low detection accuracy, poor robustness, occlusion, missed detections, and false detections. To address this issue, a multi-scale object detection algorithm based on an improved YOLOv8 has been proposed. Firstly, a lightweight attention mechanism, Triplet Attention, is introduced to enhance the algorithm’s ability to extract multi-dimensional and multi-scale features, thereby improving the receptive capability of the feature maps. Secondly, the Diverse Branch Block (DBB) is integrated into the CSP Bottleneck with two Convolutions (C2F) module to strengthen the fusion of semantic information across different layers. Thirdly, a new decoupled detection head is proposed by redesigning the original network head based on the Diverse Branch Block module to improve detection accuracy and reduce missed and false detections. Finally, the Minimum Point Distance based Intersection-over-Union (MPDIoU) is used to replace the original YOLOv8 Complete Intersection-over-Union (CIoU) to accelerate the network’s training convergence. Comparative experiments and dehazing pre-processing tests were conducted on the RTTS and VOC-Fog datasets. Compared to the baseline YOLOv8 model, the improved algorithm achieved mean Average Precision (mAP) improvements of 4.6% and 3.8%, respectively. After defogging pre-processing, the mAP increased by 5.3% and 4.4%, respectively. The experimental results demonstrate that the improved algorithm exhibits high practicality and effectiveness in foggy traffic scenarios.

关键词： Deep learning object detection foggy scenes traffic detection YOLOv8

来源：评论

学校读者我要写书评

暂无评论

Robust copy-move detection and localization of digital audio based CFCC feature

引用

Multimedia Tools and Applications 2025年第11期84卷 9573-9589页

作者： Wang, Dongyu Li, Xiaojie Shi, Canghong Niu, Xianhua Xiong, Ling Wu, Hanzhou Qian, Qing Qi, Chao School of Computer and Software Engineering Xihua University Chengdu610039 China The College of Computer Science Chengdu University of Information Technology Chengdu610225 China School of Information Shanghai University Shanghai200444 China School of Information Guizhou University of Finance and Economics Guiyang550000 China

Copy-move forgery is a common audio tampering technique in which users copy the contents of one speech and paste them into another region of the same speech signal, thus achieving the effect of tampering with the semantics. To verify the authenticity of the audio, this paper proposes a method to detect and localize audio copy-move forgery based on the cochlear filter of cochlear filter cepstral coefficients (CFCC) feature. The pitch tracking algorithm is used to distinguish the voiced and unvoiced segments in the audio, and then the CFCC coefficients are extracted for each voiced segment. The CFCC feature simulates the entire transmission process of signals in the cochlear basilar membrane using wavelet transformation. Finally, we use Pearson correlation coefficients (PCCs) and dynamic time warping (DTW) in combination to compare the similarity of voiced segments, accurately determining the tampered locations in the audio through threshold judgment. Through extensive experiments on relevant datasets, this algorithm achieves a precision rate of 98.39% and a recall rate of 98.00% in detecting audio without post-processing. Even when detecting audio that has undergone different post-processing, the precision and recall rates remain above 90%. Compared to existing methods, this approach not only achieves precise localization of replicated segments but also demonstrates superior experimental results. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

An overview on IRS-enabled sensing and communications for 6G: architectures, fundamental limits, and joint beamforming designs

引用

Science China(information Sciences) 2025年第5期68卷 170-193页

作者： Xianxin SONG Yuan FANG Feng WANG Zixiang REN Xianghao YU Ye ZHANG Fan LIU Jie XU Derrick Wing Kwan NG Rui ZHANG Shuguang CUI School of Science and Engineering (SSE) Shenzhen Future Network of Intelligence Institute (FNii-Shenzhen)and Guangdong Provincial Key Laboratory of Future Networks of Intelligence The Chinese University of Hong Kong Department of Electrical Engineering City University of Hong Kong School of Information Engineering Guangdong University of Technology Key Laboratory of Wireless-Optical Communications Chinese Academy of Sciences School of Information Science and Technology University of Science and Technology of China Computer School Beijing Information Science and Technology University National Mobile Communications Research Laboratory School of Information Science and Engineering Southeast University School of Electrical Engineering and Telecommunications University of New South Wales School of Science and Engineering Shenzhen Research Institute of Big Data The Chinese University of Hong Kong Department of Electrical and Computer Engineering National University of Singapore

This study presents an overview on intelligent reflecting surface(IRS)-enabled sensing and communication for the forthcoming sixth-generation(6G) wireless networks, in which IRSs are strategically deployed to proactively reconfigure wireless environments to improve both sensing and communication(S&C) performance. First, we exploit a single IRS to enable wireless sensing in the base station's(BS's) non-line-of-sight(NLoS) area. In particular, we present three IRS-enabled NLoS target sensing architectures with fully-passive, semi-passive, and active IRSs, respectively. We compare their pros and cons by analyzing the fundamental sensing performance limits for target detection and parameter estimation. Next, we consider a single IRS to facilitate integrated sensing and communication(ISAC), in which the transmit signals at the BS are used for achieving both S&C functionalities, aided by the IRS through reflective beamforming. We present joint transmit signal and receiver processing designs for realizing efficient ISAC, and jointly optimize the transmit beamforming at the BS and reflective beamforming at the IRS to balance the fundamental performance tradeoff between S&C. Furthermore, we discuss multi-IRS networked ISAC, by particularly focusing on multi-IRS-enabled multi-link ISAC, multi-region ISAC, and ISAC signal routing, respectively. Finally, we highlight various promising research topics in this area to motivate future work.

关键词： integrated sensing and communication (ISAC) intelligent reflecting surface (IRS) non-line-of-sight (NLoS) sensing sensing and communication tradeoff

来源：评论

学校读者我要写书评

暂无评论

Classifying distinct emotions from parents of ASD child using EEG source data by combining Bernoulli–Laplace Prior and graph neural networks

引用

Neural Computing and Applications 2025年第12期37卷 7877-7895页

作者： ArulDass, Stephen Dass Jayagopal, Prabhu School of Computer Science Engineering and Information Systems Vellore Institute of Technology Tamil Nadu Vellore632014 India

Emotion recognition using biological brain signals needs to be reliable to attain effective signal processing and feature extraction techniques. The impact of emotions in interpretations, conversations, and decision-making, has made automatic emotion recognition and examination of a significant feature in the field of psychiatric disease treatment and cure. The problem arises from the limited spatial resolution of EEG recorders. Predetermined quantities of electroencephalography (EEG) channels are used by existing algorithms, which combine several methods to extract significant data. The major intention of this study was to focus on enhancing the efficiency of recognizing emotions using signals from the brain through an experimental, adaptive selective channel selection approach that recognizes that brain function shows distinctive behaviors that vary from one individual to another individual and from one state of emotions to another. We apply a Bernoulli–Laplace-based Bayesian model to map each emotion from the scalp senses to brain sources to resolve this issue of emotion mapping. The standard low-resolution electromagnetic tomography (sLORETA) technique is employed to instantiate the source signals. We employed a progressive graph convolutional neural network (PG-CNN) to identify the sources of the suggested localization model and the emotional EEG as the main graph nodes. In this study, the proposed framework uses a PG-CNN adjacency matrix to express the connectivity between the EEG source signals and the matrix. Research on an EEG dataset of parents of an ASD (autism spectrum disorder) child has been utilized to investigate the ways of parenting of the child's mother and father. We engage with identifying the personality of parental behaviors when regulating the child and supervising his or her daily activities. These recorded datasets incorporated by the proposed method identify five emotions from brain source modeling, which significantly improves the accurac

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

Metric Network for E-Nose Drift Compensation: Few-Shot Learning for Robust Gas Sensing

引用

IEEE Sensors Journal 2025年第9期25卷 16489-16500页

作者： Tian, Yao Jiang, Qingming Zeng, Yan Peng, Linlong Sun, Jinlong Jia, Pengfei Guangxi University School of Computer and Electronic Information Nanning530004 China Guangxi University School of Electrical Engineering Nanning530004 China

This study introduces the metric drift compensation network (MDCN) to address the issue of sensor drift in electronic noses (E-noses). E-noses mimic the olfactory sense of mammals to detect odors. Sensor drift, which refers to the change in sensor outputs over time, poses a significant challenge to the reliability of E-noses. MDCN utilizes metric learning and few-shot learning (FSL) within a metric learning framework to enhance stability against drift. Its advantage lies in maintaining good classification performance even when there are few samples in the target domain or when new categories emerge in the target domain. We evaluated the performance of MDCN in scenarios of category symmetry (where the source and target domains share the same categories) and category asymmetry (where there are fewer categories in the source domain) on two datasets: the pure gas dataset and the mixed gas dataset. In category symmetry scenarios, MDCN outperformed traditional and advanced methods, demonstrating high accuracy with a minimal number of reference samples. In category asymmetry scenarios, it also showed strong generalization capabilities and high accuracy. Comprehensive ablation experiments were also conducted to prove the rationality of the model architecture and its nondependence on a large number of target domain samples. In addition, tests have shown that the model has good device transferability. © 2001-2012 IEEE.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

Optimization Design of Output Combining Network Using Irregular Structure for Broadband DPA

引用

IEEE Microwave and Wireless Technology Letters 2025年第4期35卷 480-483页

作者： Zhang, Heng Xia, Jing Ni, Zhongpeng Ge, Xiaoshuai Kong, Wa Zhang, Wence Yu, Chao Zhu, Xiao-Wei Jiangsu University School of Computer Science and Communication Engineering Zhenjiang212013 China Southeast University School of Information Science and Engineering Nanjing210096 China

This letter presents an optimization design method for broadband Doherty power amplifier (DPA) using an irregular output combining network (OCN). First, the postmatching network and the output matching networks are considered as the overall OCN, and the three-port S-parameters of the OCN are analyzed and calculated using the impedances at power back-off (PBO) and saturation. Subsequently, to improve the feasibility of optimization, several impedances that meet the requirements are used to calculate feasible S-parameter solution sets, which can be used as the optimization target. Finally, an optimization method of the OCN using an irregular structure is proposed. For verification, a 1.2-2.8-GHz broadband DPA is designed and measured. The results show that with the saturated output power exceeding 43.1 dBm, drain efficiencies of 54.1%-74.2% and 50.9%-56.5% are obtained at saturation and 6-dB PBO, respectively. © 2023 IEEE.

关键词： Doherty amplifiers

来源：评论

学校读者我要写书评

暂无评论

CBIR: a novel identification approach for college students in need based on consumer behavior psychology theory

引用

Neural Computing and Applications 2025年第6期37卷 4663-4677页

作者： Liu, Xinze Liu, Shixi Hu, Xiaojing Zhang, Yudong Fang, Xianwen School of Computer and Information Engineering Chuzhou University Chuzhou239000 China School of Computer Science and Engineering Southeast University Nanjing210096 China School of Mathematics and Big Data Anhui University of Science and Technology Huainan232001 China

The accurate identification of students in need is crucial for governments and colleges to allocate resources more effectively and enhance social equity and educational fairness. Existing approaches to identifying students in need rely on manual operations that include manually extracting consumption behavior information, statistical consumption characteristics and principal component analysis. However, this issue may lead to low prediction accuracy and inefficiency in identifying students in need. We design a three-stage framework to accurately identify college students in need from the perspective of consumer behavior psychology. The consumption behavior information is first obtained from the student consumption records using the consumption behavior clustering approach. The consumption behavior matrix is then built by extracting consumption and spatiotemporal information in different periods. Finally, a novel consumption behavior identification ResNeSt (CBIR) model is proposed to identify college students in need accurately. The experimental results on real datasets show that the CBIR model has higher prediction accuracy than the baseline models. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Social psychology

来源：评论

学校读者我要写书评

暂无评论

Foundation models for topic modeling: a case study

引用

Frontiers of computer Science 2025年第2期19卷 129-131页

作者： Han ZENG Jia-Ming SUN Chun-Shu LI Zhuying LI Tong WEI School of Computer Science and Engineering Southeast UniversityNanjing 210096China Key Laboratory of Computer Network and Information Integration Southeast UniversityNanjing 210096China

1 Introduction In Natural Language Processing(NLP),topic modeling is a class of methods used to analyze and explore textual corpora,i.e.,to discover the underlying topic structures from text and assign text pieces to different *** NLP,a topic means a set of relevant words appearing together in a particular pattern,representing some specific *** is beneficial for tracking social media trends,constructing knowledge graphs,and analyzing writing *** modeling has always been an area of extensive research in *** methods like Latent Semantic Analysis(LSA)and Latent Dirichlet Allocation(LDA),based on the“bag of words”(BoW)model,often fail to grasp the semantic nuances of the text,making them less effective in contexts involving polysemy or data noise,especially when the amount of data is small.

关键词： words semantic textual

来源：评论

学校读者我要写书评

暂无评论

Joint Watermarking and Encryption for Social Image Sharing

引用

computers, Materials & Continua 2025年第5期83卷 2927-2946页

作者： Conghuan Ye Shenglong Tan Shi Li Jun Wang Qiankun Zuo Bing Xiong School of Information Engineering Hubei University of EconomicsWuhan430205China School of Computer and Communication Engineering Changsha University of Science&TechnologyChangsha410114China

With the fast development of multimedia social platforms,content dissemination on social media platforms is becomingmore *** image sharing can also raise privacy *** encryption can protect social ***,most existing image protection methods cannot be applied to multimedia social platforms because of encryption in the spatial *** this work,the authors propose a secure social image-sharing method with watermarking/fingerprinting and ***,the fingerprint code with a hierarchical community structure is designed based on social network ***,discrete wavelet transform(DWT)from block discrete cosine transform(DCT)directly is *** that,all codeword segments are embedded into the LL,LH,and HL subbands,*** selected subbands are confused based on Game of Life(GoL),and then all subbands are diffused with singular value decomposition(SVD).Experimental results and security analysis demonstrate the security,invisibility,and robustness of our ***,the superiority of the technique is elaborated through comparison with some related image security *** solution not only performs the fast transformation from block DCT to one-level DWT but also protects users’privacy in multimedia social *** the proposed method,JPEG image secure sharing in multimedia social platforms can be ensured.

关键词： Multimedia security digital watermarking image encryption image sharing privacy protection

来源：评论

学校读者我要写书评

暂无评论

Cardinality estimation for property graph queries with gated learning approach on the graph database

引用

Multimedia Tools and Applications 2025年第11期84卷 9159-9183页

作者： He, Zhenzhen Yu, Jiong Du, Xusheng Guo, Binglei Li, Ziyang Li, Zhe School of Information Science and Engineering Xinjiang University Urumqi China School of Computer Engineering Hubei University of Arts and Science Xiangyang China School of Software Xinjiang University Urumqi China Department of Electronic and Information Engineering The Hong Kong Polytechnic University Hong Kong

With the increasing complexity of graph query processing tasks, it is difficult for users to obtain the accurate cardinality before or during the execution of query tasks. Accurate estimate query cardinality is crucial for property graph data model, which usually involves entities, multiple joins, and various types of properties (key-value pairs). While learning-based approaches have been used in query optimization, estimating the cardinality of property graph queries is still particularly challenging. In this paper, we first formally represent each property graph query and then divide the estimation process into three phases: query execution, training data generation, and model training. Secondly, we construct a data pool to deal the updating query workloads in the training data generation phase. Thirdly, we utilize the gate mechanism to develop a cardinality estimation framework based on deep learning neural networks for property graph queries in the Neo4j database, and we propose a hybrid loss function to optimize the training process. Finally, we adopt a method of pre-aggregating the underlying data to speed up query execution in the first phase. The result of the experiment on two real-world datasets shows that our optimized methods can effectively improve the estimation accuracy of the deep learning model, and our query estimator outperforms other deep learning models compared in terms of estimation accuracy. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.

关键词： Query processing

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：