检索结果-内蒙古大学图书馆

10th Symposium on Novel Optoelectronic Detection technology and Applications

作者： Fang, Xiangyi Ye, Yun Chen, Enguo College of Physics and Information Engineering Fuzhou University Fujian350108 China Fujian Science & Technology Innovation Laboratory for Optoelectronic Information of China Fuzhou350108 China

ISBN: (数字)9781510688155

ISBN: (纸本)9781510688148

A Gaussian beam is often used as the source for Bessel beam generation;however, its Gaussian intensity profile results in a tapered edge along the propagation direction. To address this issue, a flat-top beam is employed to generate a uniform intensity profile, theoretically achieving an ideal working distance in non-diffracting zones. The flat-top beam is shaped from a point light source, expanded by a convex lens, and then homogenized by an aspherical lens to achieve a flat-top beam with an evenness rate of 93%. The achievement of a long working distance has been demonstrated through simulations and experiments. Simulation results indicate that, for the same beam width, the flat-top beam generates a 33% longer working distance compared to a Gaussian beam and exhibits a "steep edge" along the propagation axis. Experimentally, the method extended the application to a remarkable 2.4 mm, compared to a Gaussian-generated Bessel beam with a full width at half maximum (FWHM) of about 1.8 mm. A beam profiler also detected a sharper intensity edge, sharpening the FWHM edge width to 0.1 mm. After optimizing details such as refining the axicon vertex angle to reduce intensity ripple along the propagation axis and smoothing the aspherical lens surface to eliminate mid-frequency interference, this solution was applied to a laser system for cutting float glass, operated by a 355 nm picosecond solid-state laser. A high-quality ultraviolet (UV) Bessel beam with a cutting depth of 2.2 mm was achieved, and the sharp intensity profile optimized the cross-section roughness. Detected by a 3D profilometer, the minimum surface roughness was reduced from 3.32 μm to 1.94 μm. This method shows promising prospects for efficient and high-quality optical glass cutting. © 2025 SPIE.

关键词： Optical glass

来源：评论

学校读者我要写书评

暂无评论

A Method for Image Super-Resolution Reconstruction with Attention Mechanism in Generative Adversarial Networks 24

A Method for Image Super-Resolution Reconstruction with Atte...

引用

13th International Conference on Computing and Pattern Recognition, ICCPR 2024

作者： Wang, Sirong Liu, Pengyu Chen, Shanji Jia, Shaohua Dong, Min Zheng, Tianyang School of Information Science and Technology Beijing University of Technology Beijing China Advanced Information Network Beijing Laboratory Beijing University of Technology Beijing China College of Physics and Electronic Information Engineering Qinghai Minzu University Qinghai China

ISBN: (纸本)9798400717482

The generation of image texture details is a crucial aspect in the assessment of the efficacy of image super-resolution reconstruction. In order to enhance the super-resolution effect of real-world degraded images, this paper proposes a method for image super-resolution reconstruction with an attention mechanism in generative adversarial networks. Firstly, a multi-angle degradation modeling process is introduced for real-world image degradation scenarios to better simulate complex real-world degradation;secondly, the structure of the Residual-in-Residual Dense Block (RRDB) commonly used in super-resolution reconstruction generative adversarial networks was optimized, and combined with the dynamic attention mechanism, a Residual-in-Residual Dense Block with Dynamic Channel and Spatial Attention Mechanism (RRDB-DCSA) is proposed, and this module is used to construct a generator for Generative adversarial networks (GAN);and lastly, locally discriminative learning loss is introduced to reduce the artifacts caused by the GAN network with the aim of learning features associated with image degradation. The experimental results demonstrate that the proposed algorithm is capable of reconstructing the detailed components of an image while simultaneously suppressing the artifacts present in scenes featuring people, animals, and landscapes. In terms of subjective assessment, the algorithm exhibits superior performance in this regard, with an improvement of nearly 17 in PSNR and SSIM, respectively. Additionally, the algorithm exhibits a reduction of nearly 15 in LPIPS and NIQE when compared with traditional and classical algorithms of the GAN network. © 2024 Copyright held by the owner/author(s).

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Real-Time Road Damage Detection Using an Optimized YOLOv9s-Fusion in IoT Infrastructure

引用

IEEE Internet of Things Journal 2025年第11期12卷 17649-17660页

作者： Waseem Khan, Muhammad Obaidat, Mohammad S. Mahmood, Khalid Sadoun, Balqies Sanaullah Badar, Hafiz Muhammad Gao, Wu Center for Bio-inspired System-on-Chip School of Computer Science Northwestern Polytechnical University Xi’an China Sahiwal Campus Pakistan The King Abdullah II School of Information Technology The University of Jordan Amman11942 Jordan The School of Communication and Computing Engineering University of Science and Technology Beijing Beijing100083 China The Amity School of Engineering and Technology Amity University Uttar Pradesh Noida201301 India The Graduate School of Intelligent Data Science National Yunlin University of Science and Technology Douliu Yunlin64002 Taiwan School of Computer and Information Engineering Henan University Kaifeng475001 China The College of Engineering Al-Balqa’ Applied University Al-Salt19117 Jordan

In IoT-enabled smart infrastructure, accurate and real-time road damage detection is crucial for enhancing road safety and optimizing maintenance processes. However, detecting road damage in complex and dynamic environments presents significant challenges, such as varying lighting conditions, diverse damage types, and the need for fast processing to enable real-time decision-making. This study introduces an advanced approach utilizing the YOLOv9s-Fusion model to overcome these challenges. Leveraging the RDD2022 dataset, which comprises 1,976 annotated images of road damage from China, we employ comprehensive data preprocessing to create optimal conditions for model training. The YOLOv9s-Fusion model integrates innovative features, including a Transformer-based auxiliary module and enhanced feature extraction layers, specifically designed to detect fine-grained damage patterns accurately. Experimental results demonstrate that the model outperforms existing approaches, achieving notable improvements in mean Average Precision (mAP) and F1-score. Ablation studies further validate the impact of our modifications, highlighting the model’s robustness in real-time detection across diverse conditions. This IoT-centric approach sets a new standard for autonomous road damage detection, significantly advancing vehicle navigation and smart infrastructure management capabilities. © 2014 IEEE.

关键词： Intelligent systems

来源：评论

学校读者我要写书评

暂无评论

An FPGA-based Key-Switching Accelerator with Ultra-High Throughput for FHE 24

An FPGA-based Key-Switching Accelerator with Ultra-High Thro...

引用

43rd International Conference on Computer-Aided Design, ICCAD 2024

作者： Lu, Zhaojun Xu, Peng Wang, Yijie Yang, Yifan Chen, Qidong Yu, Weizong Qu, Gang Hubei Key Laboratory of Distributed System Security School of Cyber Science and Engineering Huazhong University of Science and Technology China Wuhan JinYinHu Laboratory China University of Maryland College Park United States

ISBN: (纸本)9798400710773

Fully Homomorphic Encryption (FHE) enables computations directly on encrypted numbers, thereby preserving the privacy of sensitive information even in untrusted environments. However, the substantial computational overhead associated with homomorphic evaluations restricts the practical application of FHE schemes. To deal with the performance challenges, this paper proposes a hardware/software pipeline framework with a three-level cache architecture to accelerate the costly Key-Switching operation in FHE. This framework supports the dynamically reconfigurable processing mode, two parallelism strategies, and flexible control flow, effectively breaking the compute-bound and the memory-bound limitations. A no-stall and conflict-free memory mapping algorithm is implemented on the Xilinx U55C FPGA platform that boosts the throughput of ciphertext-ciphertext multiplication to 395 operations per second, which 1.4× and 2.5× speeds up the FPGA-based (HPCA’23) and GPU-based (HPCA’23) schemes with the same parameter set and precision. © 2024 Copyright is held by the owner/author(s).

关键词： Sensitive data

来源：评论

学校读者我要写书评

暂无评论

Frontal midline theta power accounts for inter-individual differences in motor learning ability

引用

Experimental Brain Research 2025年第6期243卷 1-14页

作者： Fukuda, Yuya Uehara, Kazumasa Neural Information Dynamics Laboratory Department of Computer Science and Engineering Toyohashi University of Technology 1-1 Hibarigaoka Tempaku-cho Aichi Toyohashi Japan

Recent neurophysiological studies have demonstrated that frontal midline theta (FMT) activity plays a significant role in motor learning. One of the key challenges in motor learning is to understand the interindividual variability in learning proficiency rates, yet the underlying neural mechanisms remain unclear. To address this open question, this study recorded electroencephalogram activity from twenty-one healthy participants during a visuomotor tracking task to investigate whether modulation of FMT power and the theta phase synchronization across trials (theta phase consistency) during motor preparation could explain individual differences in learning proficiency. We found a significant positive correlation between increased FMT power during motor preparation and learning proficiency rates. Specifically, individuals with greater FMT power exhibited faster learning rates. In contrast, no significant correlation was observed between the consistency of the theta phase during motor preparation and learning proficiency. Together, these findings highlight that the FMT power, rather than phase synchrony, is closely associated with motor learning efficiency. This study provides a novel perspective for understanding the causes of individual differences in motor learning and further corroborates the previous evidence showing FMT power contributes to motor learning processes. © The Author(s) 2025.

关键词： EEG Frontal midline Theta Individual differences Motor learning Phase oscillation

来源：评论

学校读者我要写书评

暂无评论

CRFENet: Cross-Spectral Stereo Matching Network Exploiting Reciprocal Feature Evolution 2nd

CRFENet: Cross-Spectral Stereo Matching Network Exploiting R...

引用

2nd International Conference on the Frontiers of Robotics and Software engineering, FRSE 2024

作者： Li, Yangqin Zhang, Guanghui Zhu, Dongchen Shi, Wenjun Wang, Lei Zhang, Xiaolin Shanghai Institute of Microsystem and Information Technology Bionic Vision System Laboratory State Key Laboratory of Transducer Technology Chinese Academy of Sciences Shanghai200050 China University of Chinese Academy of Sciences Beijing100049 China School of Information Science and Technology ShanghaiTech University Shanghai201210 China

ISBN: (纸本)9789819630127

Cross-spectral stereo matching opens up a new route for disparity estimation in low-light scenes. However, the variations in light absorption and reflection properties for objects at different wavelengths create substantial domain gaps in cross-spectral image pairs. Accurately matching pixel pairs with significant domain gaps remains challenging. In this work, we propose a novel lightweight cross-spectral stereo matching network, dubbed CRFENet, exploiting reciprocal feature evolution. It enables the complementary evolution of the disparity prediction network and spectral translation network. To facilitate feature interactions between different domains to improve disparity results in regions with significant domain gaps, we design a domain mutual rail (DMR) module, which integrates monocular spectral and binocular cross-spectral features. Furthermore, we present a domain gap consistency constraint and corresponding confidence-weighted loss function, which aim to alleviate the matching challenge in regions with significant domain gaps through decoupling domain gap and disparity. Experimental results indicate that our CRFENet achieves state-of-the-art performance on the PittsStereo RGB-NIR and DrivingStereo datasets, especially for challenging regions with significant domain gaps. Besides, the model parameters are only 3.6M, which means high practicality. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Stereocenters

来源：评论

学校读者我要写书评

暂无评论

Modular access toπ-conjugated planar chiral diradicaloids using redox-active pillararenes

引用

science China Chemistry 2025年

作者： Yuting Xue Yi Zeng Yafei Shi Yu Tian Cuiping Sun Niu Zhang Xiaoyan Zheng Pangkuan Chen Beijing Key Laboratory of Photoelectronic/Electrophotonic Conversion Materials Key Laboratory of Cluster Science of the Ministry of Education Key Laboratory of Medical Molecule Science and Pharmaceutics Engineering of the Ministry of Industry and Information Technology School of Chemistry and Chemical Engineering Beijing Institute of Technology Analysis & Testing Centre Beijing Institute of Technology

Open-shell chiral diradicals that feature both the unique chiroptical properties and tunable spin states have kindled numerous interests in organic synthesis and material *** present a new class of such model systems based on theπ-extended planar chiral pillar[5]arene derivatives P5Flu and P5Cz,where the pillar[5]arene-based chiral scaffolds are coupled with sterically encumbered fluorenyl and carbazolyl moieties as highly luminescent redox-active emitters,*** molecules in neutral states can readily undergo chemical oxidations leading toπ-conjugated chiral diradicals of the dehydrogenated P5Flu2·and a positively charged dication P5Cz2+2·.Characterization of the resulting Müller’s open-shell species by variable-temperature electron spin resonance （VT-ESR）,UV-vis-NIR absorption,VT-NMR and by computational studies revealed a singlet (ΔES-T=?1.63 kcal mol-1) and a nearly degenerate singlet-triplet ground state (ΔES-T=0.05 kcal mol-1,with a diradical character index y0=0.89) for P5Cz2+2·and P5Flu2·,*** particularly show in this work a new journey for pillar[5]arenes from the conventionally available closed-shell structures to different open-shell spin states with potential applications in chirospintronics and NIR-active chiroptoelectronic materials.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Color Decoupling for Multi-Illumination Color Constancy

引用

IEEE Transactions on Circuits and systems for Video technology 2025年第5期35卷 4087-4099页

作者： Zhang, Wen Tan, Zhenshan Zhang, Li Li, Zhijiang Wuhan University School of Information Management Research Center of Graphic Communication Printing and Packaging Wuhan430079 China Nanjing University of Information Science and Technology School of Computer Science Nanjing210044 China Wuhan University State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing Wuhan430079 China Wuhan University School of Information Management Wuhan430079 China

Current multi-illumination color constancy methods typically estimate illumination for each pixel directly. However, according to the multi-illumination imaging equation, the color of each pixel is determined by various components, including the innate color of the scene content, the colors of multiple illuminations, and the weightings of these illuminations. Failing to distinguish between these components results in color coupling. On the one hand, there is color coupling between illumination and scene content, where estimations are easily misled by the colors of the content, and the distribution of the estimated illuminations is relatively scattered. On the other hand, there is color coupling between illuminations, where estimations are susceptible to interference from high-frequency and heterogeneous illumination colors, and the local contrast is low. To address color coupling, we propose a Color Decoupling Network (CDNet) that includes a Content Color Awareness Module (CCAM) and a Contrast HArmonization Module (CHAM). CCAM learns scene content color priors, decoupling the colors of content and illuminations by providing the model with the color features of the content, thereby reducing out-of-gamut estimations and enhancing consistency. CHAM constrains feature representation, decoupling illuminants by mutual calibration between adjacent features. CHAM utilizes spatial correlation to make the model more sensitive to the relationships between neighboring features and utilizes illumination disparity degree to guide feature classification. By enhancing the uniqueness of homogeneous illumination features and the distinctiveness of heterogeneous illumination features, CHAM improves local edge contrast. Additionally, by allocating fine-grained margin coefficients to emphasize the soft distinctiveness of similar illumination features, further enhancing local contrast. Extensive experiments on single- and multi-illumination benchmark datasets demonstrate that the propose

关键词： Calibration

来源：评论

学校读者我要写书评

暂无评论

Improving Speaker Verification Back-End with Graph Neural Networks

引用

Journal of Shanghai Jiaotong University (science) 2025年 1-9页

作者： Chen, Jinfeng Fang, Zhihua He, Liang School of Computer Science and Technology Xinjiang University Urumqi830017 China Xinjiang Key Laboratory of Signal Detection and Processing Urumqi830017 China School of Intelligence Science and Technology Xinjiang University Urumqi830017 China Department of Electronic Engineering Beijing National Research Center for Information Science and Technology Tsinghua University Beijing100084 China

Currently, research on speaker verification tasks is primarily concentrated on enhancing deep speaker models to extract high-quality speaker embeddings. Nevertheless, this speaker embeddings can be regarded as potential graph structures. Therefore, this paper introduces a graph neural network-based speaker verification back-end approach, treating the speaker embeddings derived from the front-end as graph structures and leveraging graph neural networks to uncover the relationships among embeddings, thereby achieving superior-quality speaker embeddings. In addition, we propose a group updating method to solve the problem of memory overflow when the number of nodes is excessive. Extensive experiments and ablation studies were carried out on the VoxCeleb dataset, and the experimental outcomes validate the effectiveness of our proposed graph neural network speaker back-end in significantly boosting the performance of speaker verification systems. (Figure presented.) © Shanghai Jiao Tong University 2025.

关键词： Graph embeddings

来源：评论

学校读者我要写书评

暂无评论

High Spatio-Temporal Imaging Reconstruction for Hybrid Spike-RGB Cameras

IEEE Transactions on Computational Imaging

引用

IEEE Transactions on Computational Imaging 2025年 11卷 586-598页

作者： Xia, Lujie Xiong, Ruiqin Zhao, Jing Wang, Lizhi Zhu, Shuyuan Fan, Xiaopeng Huang, Tiejun Peking University National Engineering Research Center of Visual Technology School of Computer Science Beijing100871 China Beijing Normal University School of Artificial Intelligence Beijing100875 China University of Electronic Science and Technology of China School of Information and Communication Engineering Chengdu611731 China Department of Computer Science and Technology Harbin150001 China Peng Cheng Laboratory Shenzhen518055 China

The acquisition of high-resolution image sequence for dynamic scenes of fast motion remains challenging due to motion blur caused by fast object movement. As a novel neuromorphic sensor, spike camera records the changing light intensity via spike stream of ultra-high temporal resolution, excelling in motion recording but limited in spatial resolution. This paper proposes a method for high spatio-temporal resolution (HSTR) imaging with a hybrid Spike-RGB camera, utilizing the information from spike stream to enhance the temporal resolution and the information from RGB images to enhance the spatial resolution of texture details. For this purpose, we present HSTR-Net, a dedicated network to process the spike and RGB data, which incorporates three key innovations: 1) A temporal control encoder enabling flexible temporal reconstruction through spike stream processing with embedded time parameters, eliminating the requirement to train multiple inference models;2) Motion-aware feature projection that aligns RGB frame details to target timestamps using spike-derived motion offsets;3) An adaptive transformer-based fusion strategy establishing cross-modal spatial correlations through mutual attention mechanisms. Extensive experiments demonstrate state-of-the-art performance on synthetic benchmark datasets with 5.23 dB PSNR and 6.94% SSIM improvement. It also shows visually impressive performance on real-world captured spike dataset. © 2015 IEEE.

关键词： High speed cameras

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：