检索结果-内蒙古大学图书馆

TTST: A Top-k Token Selective Transformer for Remote Sensing Image Super-Resolution

IEEE Transactions on Image Processing 2024年 33卷 738-752页

作者： Xiao, Yi Yuan, Qiangqiang Jiang, Kui He, Jiang Lin, Chia-Wen Zhang, Liangpei Wuhan University School of Geodesy and Geomatics Wuhan430079 China Harbin Institute of Technology School of Computer Science and Technology Harbin150001 China National Tsing Hua University Department of Electrical Engineering The Institute of Communications Engineering Hsinchu300 Taiwan Wuhan University State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing Wuhan430079 China

Transformer-based method has demonstrated promising performance in image super-resolution tasks, due to its long-range and global aggregation capability. However, the existing Transformer brings two critical challenges for applying it in large-area earth observation scenes: (1) redundant token representation due to most irrelevant tokens;(2) single-scale representation which ignores scale correlation modeling of similar ground observation targets. To this end, this paper proposes to adaptively eliminate the interference of irreverent tokens for a more compact self-attention calculation. Specifically, we devise a Residual Token Selective Group (RTSG) to grasp the most crucial token by dynamically selecting the top-k keys in terms of score ranking for each query. For better feature aggregation, a Multi-scale Feed-forward Layer (MFL) is developed to generate an enriched representation of multi-scale feature mixtures during feed-forward process. Moreover, we also proposed a Global Context Attention (GCA) to fully explore the most informative components, thus introducing more inductive bias to the RTSG for an accurate reconstruction. In particular, multiple cascaded RTSGs form our final Top-k Token Selective Transformer (TTST) to achieve progressive representation. Extensive experiments on simulated and real-world remote sensing datasets demonstrate our TTST could perform favorably against state-of-the-art CNN-based and Transformer-based methods, both qualitatively and quantitatively. In brief, TTST outperforms the state-of-the-art approach (HAT-L) in terms of PSNR by 0.14 dB on average, but only accounts for 47.26% and 46.97% of its computational cost and parameters. The code and pre-trained TTST will be available at https://***/XY-boy/TTST for validation. © 1992-2012 IEEE.

关键词： Remote sensing

来源：评论

学校读者我要写书评

暂无评论

Local-Global Temporal Difference Learning for Satellite Video Super-Resolution

引用

IEEE Transactions on Circuits and Systems for Video Technology 2024年第4期34卷 2789-2802页

作者： Xiao, Yi Yuan, Qiangqiang Jiang, Kui Jin, Xianyu He, Jiang Zhang, Liangpei Lin, Chia-Wen The School of Geodesy and Geomatics Wuhan University Wuhan430000 China The School of Computer Science and Technology Harbin Institute of Technology Harbin150000 China The State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing Wuhan University Wuhan430000 China The Department of Electrical Engineering the Institute of Communications Engineering National Tsing Hua University Hsinchu30013 Taiwan

Optical-flow-based and kernel-based approaches have been extensively explored for temporal compensation in satellite Video Super-Resolution (VSR). However, these techniques are less generalized in large-scale or complex scenarios, especially in satellite videos. In this paper, we propose to exploit the well-defined temporal difference for efficient and effective temporal compensation. To fully utilize the local and global temporal information within frames, we systematically modeled the short-term and long-term temporal discrepancies since we observe that these discrepancies offer distinct and mutually complementary properties. Specifically, we devise a Short-term Temporal Difference Module (S-TDM) to extract local motion representations from RGB difference maps between adjacent frames, which yields more clues for accurate texture representation. To explore the global dependency in the entire frame sequence, a Long-term Temporal Difference Module (L-TDM) is proposed, where the differences between forward and backward segments are incorporated and activated to guide the modulation of the temporal feature, leading to a holistic global compensation. Moreover, we further propose a Difference Compensation Unit (DCU) to enrich the interaction between the spatial distribution of the target frame and temporal compensated results, which helps maintain spatial consistency while refining the features to avoid misalignment. Rigorous objective and subjective evaluations conducted across five mainstream video satellites demonstrate that our method performs favorably against state-of-the-art approaches. Code will be available at https://***/ XY-boy/LGTD. © 2023 IEEE.

关键词： Three dimensional displays

来源：评论

学校读者我要写书评

暂无评论

Successive Pose Estimation and Beam Tracking for mmWave Vehicular Communication Systems

Successive Pose Estimation and Beam Tracking for mmWave Vehi...

引用

2023 IEEE Globecom Workshops, GC Wkshps 2023

作者： Liu, Cen Zhu, Guangxu Liu, Fan Liu, Yuanwei Huang, Kaibin National University of Singapore Department of Electrical and Computer Engineering Singapore Shenzhen Research Institute of Big Data Shenzhen China Peng Cheng Laboratory Shenzhen China Guangzhou China Southern University of Science and Technology Department of Electronic and Electrical Engineering Shenzhen China School of Electronic Engineering and Computer Science Queen Mary University of London London United Kingdom The University of Hong Kong Department of Electrical and Electronic Engineering Hong Kong

ISBN: (纸本)9798350370218

The millimeter wave (mmWave) radar sensing-aided communications in vehicular mobile communication systems is investigated. To alleviate the beam training overhead under high mobility scenarios, a successive pose estimation and beam tracking (SPEBT) scheme is proposed to facilitate mmWave communications with the assistance of mmWave radar sensing. The proposed SPEBT scheme first resorts to a Fast Conservative Filtering for Efficient and Accurate Radar odometry (Fast-CFEAR) approach to estimate the vehicle pose consisting of 2-dimensional position and yaw from radar point clouds collected by mmWave radar sensor. Then, the pose estimation information is fed into an extend Kalman filter to perform beam tracking for the line-of-sight channel. Owing to the intrinsic robustness of mmWave radar sensing, the proposed SPEBT scheme is capable of operating reliably under extreme weather/illumination conditions and large-scale global navigation satellite systems (GNSS)-denied environments. The practical deployment of the SPEBT scheme is verified through rigorous testing on a real-world sensing dataset. Simulation results demonstrate that the proposed SPEBT scheme is capable of providing precise pose estimation information and accurate beam tracking output, while reducing the proportion of beam training overhead to less than 5% averagely. © 2023 IEEE.

关键词： beam tracking extended Kalman filter millimeter wave pose estimation radar odometry Sensing-aided communications

来源：评论

学校读者我要写书评

暂无评论

VISION-iT:A Framework for Digitizing Bubbles and Droplets

引用

Energy and AI 2024年第1期15卷 26-35页

作者： Youngjoon Suh Sanghyeon Chang Peter Simadiris Tiffany B.Inouye Muhammad Jahidul Hoque Siavash Khodakarami Chirag Kharangate Nenad Miljkovic Yoonjin Won Department of Mechanical and Aerospace Engineering University of CaliforniaIrvine4200 Engineering GatewayCA 92617-2700USA Department of Electrical Engineering and Computer Science University of CaliforniaIrvine5200 Engineering HallCA 92617-2700USA Department of Mechanical Science and Engineering University of Illinois at Urbana-ChampaignUrbanaIL 61801USA Mechanical and Aerospace Engineering Department Case Western Reserve UniversityClevelandOH44106USA Department of Electrical and Computer Engineering University of Illinois at Urbana-ChampaignUrbanaIL61801USA Materials Research Laboratory University of Illinois at Urbana-ChampaignUrbanaIL61801USA International Institute for Carbon Neutral Energy Research(WPI-12CNER) Kyushu University744 Moto-okaNishi-kuFukuoka 819-0395Japan

Quantifying the nucleation processes involved in liquid-vapor phase-change phenomena,while dauntingly challenging,is central in designing energy conversion and thermal management *** technological advances in the deep learning and computer vision field offer the potential for quantifying such complex two-phase nucleation processes at unprecedented *** leveraging these new technologies,a multiple object tracking framework called“vision inspired online nuclei tracker(VISION-iT)”has been proposed to extract large-scale,physical features residing within boiling and condensation ***,extracting high-quality features that can be integrated with domain knowledge requires detailed discussions that may be field-or case-specific *** this regard,we present a demonstration and discussion of the detailed construction,algorithms,and optimization of individual modules to enable adaptation of the framework to custom *** concepts and procedures outlined in this study are transferable and can benefit broader audiences dealing with similar problems.

关键词： Deep learning computer vision Nucleation Heat transfer Phase-change phenomena

来源：评论

学校读者我要写书评

暂无评论

Multi-model ensemble conformal prediction in dynamic environments 24

Multi-model ensemble conformal prediction in dynamic environ...

引用

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Erfan Hajihashemi Yanning Shen Department of Electrical Engineering & Computer Science University of California Irvine

ISBN: (纸本)9798331314385

Conformal prediction is an uncertainty quantification method that constructs a prediction set for a previously unseen datum, ensuring the true label is included with a predetermined coverage probability. Adaptive conformal prediction has been developed to address data distribution shifts in dynamic environments. However, the efficiency of prediction sets varies depending on the learning model used. Employing a single fixed model may not consistently offer the best performance in dynamic environments with unknown data distribution shifts. To address this issue, we introduce a novel adaptive conformal prediction framework, where the model used for creating prediction sets is selected 'on the fly' from multiple candidate models. The proposed algorithm is proven to achieve strongly adaptive regret over all intervals while maintaining valid coverage. Experiments on real and synthetic datasets corroborate that the proposed approach consistently yields more efficient prediction sets while maintaining valid coverage, outperforming alternative methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Modeling the electronic band-structure of strained long-wavelength Type-II superlattices using the scattering matrix method

引用

Hongwai Yu Haomibo Xuebao/Journal of Infrared and Millimeter Waves 2025年第3期44卷 345-350页

作者： Haddadi, Abbas Brown, Gail Razeghi, Manijeh Center for Quantum Devices Department of Electrical Engineering and Computer Science Northwestern University EvanstonIL60208 United States

This study introduces a comprehensive theoretical framework for accurately calculating the electronic band-structure of strained long-wavelength InAs/GaSb type-II superlattices. Utilizing an eight-band k · p Hamilto⁃ nian in conjunction with a scattering matrix method，the model effectively incorporates quantum confinement，strain effects，and interface states. This robust and numerically stable approach achieves exceptional agreement with experimental data，offering a reliable tool for analyzing and engineering the band structure of complex multi⁃ layer systems. © 2025 Chinese Optical Society. All rights reserved.

关键词： II-VI semiconductors

来源：评论

学校读者我要写书评

暂无评论

Design Drivers of an Attitude Control System for Small CubeSats Using Magnetorquers 73

Design Drivers of an Attitude Control System for Small CubeS...

引用

73rd International Astronautical Congress, IAC 2022

作者： Neagu, Catalin-Daniel Chams, Fadel Brune, Eric Friis-Liby, Linn Soria, Carmen Fuentes Graff, Erik Gustavsson, Jimmy Hestad, Theresia Golemis, Aris Grethen-Bußmann, Antonia Laufer, Rene Lindkvist, Edvin Manamon, Sorcha Mac Maneepong, Chayapol Mayer, Kyle Peroy, Cristobal Nieto Sanchez, Gabriel Cadenas, Carlos Sarille Sonna, Mrunmayee Witt, Gabrielle Widenfelt, Axel Luleå University of Technology Department of Computer Science Electrical and Space Engineering Sweden

Attitude control in the case of 1U CubeSats is a crucial design choice in any mission of such scale. This leads to extensive studies being conducted before deciding on the architecture of this subsystem. In the case of Luleå University Technology's first student-led CubeSat project, Atmospheric Polar Transmission Alignment Satellite (APTAS), the driving design choices of the Attitude Determination and Control System (ADCS) were predominately impacted by its onboard payload requirements. For successful operations of the camera, the ADCS must ensure a stability of less than 0.58 deg/s, while the pointing accuracy for both camera and EISCAT antenna system shall be less than 10 deg. In this regard, following a study of suitable hardware designs that would fulfill those requirements, the satellite shall be equipped with magnetometers, coarse sun sensors (CSS) and a gyroscope. It is widely-known that ADCS-related tasks are notoriously computationally heavy. This presents another limitation of 1U satellites: available computational load. Preliminary studies and simulation have shown that an usual algorithm in the form of a Kalman Filter (or similar applications) would interfere with payload operations due to subsequent sensor-fusion iterations and the heavy computational load resulting from it. For this reason, the TRIAD attitude determination method was deemed a fitting alternative, due to its computational efficiency and lower number of operations. All data-processing tasks must also be conducted through the one on-board computer, due to the volume constraints of 1U. In terms of control, the conducted simulations show that the B-dot algorithm for detumbling motion and the quaternion feedback control for nadir pointing are capable of meeting the requirements of the mission. For optimal payload operations, four operation modes have been defined, each to be functional at different stages of the satellite's mission. Each mode utilizes a distinct arrangement of ADCS component

关键词： Cameras

来源：评论

学校读者我要写书评

暂无评论

Optimum Estimation of Coastal Currents Using Moving Vehicles

引用

Journal of Atmospheric and Oceanic Technology 2023年第12期40卷 1431-1441页

作者： Chen, Kuanyu Huang, Chen-Fen Zheng, Zhe-Wen Lin, Sheng-Fong Liu, Jin-Yuan Guo, Jenhwa Institute of Oceanography National Taiwan University Taipei Taiwan Department of Power Vehicle and Systems Engineering National Defense University Taoyuan Taiwan Department of Earth Science National Taiwan Normal University Taipei Taiwan Department of Marine Environmental Engineering National Kaohsiung University of Science and Technology Kaohsiung Taiwan Department of Electrical and Computer Engineering Tamkang University New Taipei City Taiwan Department of Engineering Science and Ocean Engineering National Taiwan University Taipei Taiwan

Ocean acoustic tomography (OAT) deploys most moored stations on the periphery of the tomographic region to sense the solenoidal current field. Moving vehicle tomography (MVT), an advancement of OAT, not only samples the region from various angles for improving the resolution of mapped currents but also acquires information about the ir-rotational flow due to the sampling points inside the region. To reconstruct a complete two-dimensional current field, the spatial modes derived from the open-boundary modal analysis (OMA) are preferable to the conventional truncated Fourier series since the OMA technique describes the solenoidal and irrotational flows efficiently in which all modes satisfy the coastline and open boundary conditions. Comparisons of the reconstructions are presented using three different representations of currents. The first two representations explain only the solenoidal flow: the truncated Fourier series and the OMA Dirichlet modes. The third representation, accounting for the solenoidal and irrotational flows, uses all the OMA modes. For reconstructing the solenoidal flow, the OMA representation with the Dirichlet modes performs better than the Fourier series. A large difference appears near the bay mouth, where the OMA-Dirichlet reconstruction shows a better fit to the uniform currents. However, considerable uncertainty exists outside the bay mouth where the irrotational currents dominate. This can be improved by the third representation with the inclusion of the Neumann and boundary modes. The reconstruction results using field data were validated against the acoustic Doppler current profiler (ADCP) measurements. Additionally, incorporating constraints from ADCP measurements enhances the accuracy of the reconstruction. © 2023 American Meteorological Society.

关键词： Modal analysis

来源：评论

学校读者我要写书评

暂无评论

Image Transliteration from Vernacular Languages 2nd

Image Transliteration from Vernacular Languages

引用

2nd Conference on Flexible Electronics for Electric Vehicles, FlexEV 2021

作者： Gupta, Yogesh Saraswat, Amit Raghuwanshi, Ghanshyam Department of Computer Science BML Munjal University Gurugram India Department of Electrical Engineering Manipal University Jaipur Rajasthan Jaipur India Department of Computer and Communication Engineering Manipal University Jaipur Rajasthan Jaipur India

ISBN: (纸本)9789811905872

India is a multi-cultured and multi-language country, which has 23 official languages, and Hindi is known as the mother language as most of the people speak Hindi here. This paper implements a model which can extract text from images and then transliterates the text written from Indic language (Hindi) to English language. It presents distinctive issues of various text styles and patterns, comparability in structure. Therefore, it is essential to discover viable techniques for text extraction. It can be useful for a tourist to easily read the signboards or names of shops in North India as the focus of this work is to work on Hindi language. A great deal of enhancements has been made in the field deep learning and natural learning processing. This paper implements text extraction and transliteration in an efficient manner by using approaches which provide accurate results. Google Vision API and natural language processing have been used for text extraction and performing transliteration, respectively, in this paper. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Extraction

来源：评论

学校读者我要写书评

暂无评论

MilliMobile: An Autonomous Battery-free Wireless Microrobot 23

MilliMobile: An Autonomous Battery-free Wireless Microrobot

引用

29th Annual International Conference on Mobile Computing and Networking, MobiCom 2023

作者： Johnson, Kyle Englhardt, Zachary Arroyos, Vicente Yin, Dennis Patel, Shwetak Iyer, Vikram Paul G. Allen School of Computer Science and Engineering University of Washington Seattle60015481 United States Department of Electrical and Computer Engineering University of Washington Seattle60015481 United States

ISBN: (纸本)9781450399906

We present MilliMobile: a first of its kind battery-free autonomous robot capable of operating on harvested solar and RF power. We challenge the conventional assumption that motion and actuation are beyond the capabilities of battery-free devices and demonstrate completely untethered autonomous operation in realistic indoor and outdoor lighting as well as RF power delivery scenarios. We show first that through miniaturizing a robot to gram scale, we can significantly reduce the energy required to move it. Second, we develop methods to produce intermittent motion by discharging a small capacitor (47 - 150 μF) to move a motor in discrete steps, enabling motion from as little as 50 μW of power or less. We further develop software defined techniques for maximizing power harvesting. MilliMobile operates in the optimal part of the charging curve by varying the charging time to achieve maximum speeds of up to 5.5 mm/*** MilliMobile prototype has a 10x10 mm chassis and weighs less than 1.1 g. Our robot can carry payloads 3 times its own weight, and only experiences a 25% reduction in speed when carrying a 1 g payload. We demonstrate operation on 10 different surfaces ranging from wood and concrete to compact soil. We further show the ability to cold-start and move in light conditions as low as 20 W/m2 and -10 dBm of RF power. In addition to operating on harvested power, our robot demonstrates sensor and control autonomy by seeking light using onboard photodiodes, and can transmit sensor data wirelessly to a base station over 200 m away. © 2023 Owner/Author(s).

关键词： Energy harvesting

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：