检索结果-内蒙古大学图书馆

Unpaired Overwater Image Defogging Using Prior Map Guided Cycle-Consistent Generative Adversarial Network

学校读者我要写书评

暂无评论

SSRN

SSRN 2024年

作者： Mo, Yaozong Li, Chaofeng Ren, Wenqi Wang, Wenwu Wu, Xiao-Jun Shanghai201306 China School of Cyber Science and Technology Sun Yat-sen University Shenzhen518000 China Center for Vision Speech and Signal Processing Department of Electrical and Electronic Engineering University of Surrey Surrey SurreyGU2 7XH United Kingdom School of Artificial Intelligence and Computer Science Jiangnan University Wuxi 214122 China

Existing image defogging approaches have made significant advancements. But their effectiveness in addressing overwater foggy images remains limited. Current methods are predominantly optimized for land scenes, which causes poor performance when applied to overwater images, since overwater scenes typically contain large expanses of sky and water. In this paper, we propose a prior map guided cycle-consistent generative adversarial network (PG-CycleGAN) for defogging of images with overwater scenes. To promote the recovery of the objects on water in the image, two loss functions are exploited for the network where a prior map is designed to invert the dark channel and the min-max normalization is used to suppress the sky and emphasize objects. However, due to the unpaired training set, the network may learn an under-constrained domain mapping, leading to artifacts and loss of details. Thus, we propose an intuitive Upscaling Inception Module (UIM) and a Long-range Residual Coarse-to-fine framework (LRC) to mitigate this issue. Extensive experimental results demonstrate that PG-CycleGAN is effective in overwater image defogging. © 2024, The Authors. All rights reserved.

关键词： Generative adversarial networks

HILBERT-KAMKE EQUATIONS AND GEOMETRIC DESIGNS OF DEGREE FIVE FOR CLASSICAL ORTHOGONAL POLYNOMIALS

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Mishima, Teruyuki Lu, Xiao-Nan Sawa, Masanori Uchida, Yukihiro Accenture Japan Ltd 2-2-2 Nakanoshima Kita-ku Osaka530-0005 Japan Department of Electrical Electronics and Computer Engineering Faculty of Engineering Gifu University 1-1 Yanagido Gifu-shi Gifu501-1193 Japan Graduate School of System Informatics Kobe University 1-1 Rokkodai Nada Kobe Hyogo657-8501 Japan Department of Mathematical Sciences Graduate School of Science Tokyo Metropolitan University 1-1 Minami-Osawa Hachioji Tokyo192-0397 Japan

In this paper we elucidate the advantage of examining the connections between Hilbert-Kamke equations and geometric designs, or Chebyshev-type quadrature, for classical orthogonal polynomials. We first establish a classification theorem for such 5-designs with 6 points. The proof is based on an elementary polynomial identity and some advanced techniques on the computation of the genus of a certain irreducible curve. We then prove a necessary and sufficient condition for the existence of 5-designs with rational points, especially for the Chebyshev measure of the first kind. It is noteworthy that this result presents a completely explicit construction of rational designs. Moreover, we create novel connections among Hilbert-Kamke equations, geometric designs and the Prouhet-Tarry-Escott (PTE) problem. For example, we establish that the 5-designs with 6 points for the Chebyshev measure appear in the famous parametric solution for the PTE problem found by Borwein (2002).MSC Codes 33C45, 05E99, 65D32 (Primary) 12E10, 11E76 (Secondary) Copyright © 2025, The Authors. All rights reserved.

关键词： Geometry

On Matching Finger-Selfies Using Deep Scattering Networks

学校读者我要写书评

暂无评论

IEEE Transactions on Biometrics, Behavior, and Identity Scie...

IEEE Transactions on Biometrics, Behavior, and Identity science 2020年第4期2卷 350-362页

作者： Malhotra, Aakarsh Sankaran, Anush Vatsa, Mayank Singh, Richa Department of Computer Science Engineering IIIT-Delhi New Delhi110020 India Deeplite MontrealH4C 2K1 Canada Department of Cse Iit Jodhpur Jodhpur342037 India

With the advancements in technology, smartphones' capabilities have increased immensely. For instance, the smartphone cameras are being used for face and ocular biometric-based authentication. This research proposes finger-selfie based authentication mechanism, which uses a smartphone camera to acquire a selfie of a finger. In addition to personal device-level authentication, finger-selfies may also be matched with livescan fingerprints present in the legacy/national ID databases for remote or touchless authentication. We propose an algorithm which comprises segmentation, enhancement, Deep Scattering Network based feature extraction, and Random Decision Forest to authenticate finger-selfies. This paper also presents one of the largest finger-selfie database with over 19, 400 images. The images in the IIIT-D Smartphone Finger-selfie Database v2 are captured using multiple smartphones and include variations due to background, illumination, resolution, and sensors. Results and comparison with existing algorithms show the efficacy of the proposed algorithm which yields equal error rates in the range of 2.1-5.2% for different experimental protocols. © 2019 IEEE.

关键词： Smartphones

The Development Results of the Current SSPS Project and the Progress Status of the Follow-on OHISAMA Project for the Realization of the Operational SSPS

学校读者我要写书评

暂无评论

The Development Results of the Current SSPS Project and the ...

2024 IAF Space Power Symposium at the 75th International Astronautical Congress, IAC 2024

作者： Ijichi, Koichi Yanagawa, Hiroki Yanagi, Kouki Kitabatake, Hidetoslii Iiiada, Hitoini Kasliimura, Osamu Tanaka, Koji Ishikawa, Ryo Honjo, Kazuhiko Iskiimira, Kosei Ogawa, Akihiko Yamaguchi, Koji Japan Space Systems 3-5-8 Shibakoen Minato-ku Tokyo105-0011 Japan Institute of Space and Astronautical Science Japan Aerospace Exploration Agency 3-1-1 Yoshinodai Chuo-ku Kanagawa Sagamihara252-5210 Japan Department of Computer and Network Engineering The University of Electro-Communications 1-5-1 Chofugaoka C-hofu Tokyo1S2-S5S5 Japan Dept. of Modern Mechanical Engineering School of Creative Science and Engineering Waseda University 30401 Okubo Shinjuku-ku Tokyo169-S555 Japan Technologies Development Department IHI Aerospace Co. Ltd. 900 Fujiki Tomioka-shi Gunma-ken370-2398 Japan 1-7-8 Nishi-Kanagawa Kanagawa-ku Kanagawa-ken Yokohama-City221-0822 Japan

ISBN: (纸本)9798331312213

In the year of 2006, the METI of Japan established the SSPS '"USEF model"as the target SSPS configuration to be realized and also established the road map to achieve by the year 2050 based on the conclusion of the discussion based on the concept development, the related FS and also the basic technology development started from the end of 1980th by the committee members consisted of the space related universities and the organizations. Since then, the step-by-step technical development activities started and steadily proceeded according the roadmap for the realization of the ultimate carbon-free and Net-Zero power generation means. The horizontal power beaming and transmission technology development was the first step, and then proceed to the vertical transmission as the next steps. And then more, stepped into the development of the Power Generation and Transmission Panel (PGTP) which is the basic element of the USEF model, improvement of the efficiency of the DC to RP conversion efficiency and also to extend the distance of the WPT to the range of several kilo-meters. This phase stepped into the final year with full success. In parallel with current development phase but according to the road, the demonstration by 180kg class small satellite project called OHISAMA (On-orbit experiment of High-precision beam control using small SAtellite for MicrowAve power transmission) project was started to demonstrate and develop the precise power beaming technology from the LEO to the ground in the JFY of 2025. Our step-by-step approach towards the realization of USEF model according to the roadmap will be introduced with the results of the development activities, and also the progress of the OHISAMA project will be introduced and discussed. © 2024 by the International Astronautical Federation (IAF). All rights reserved.

关键词： Microwave power transmission

New antenna selection schemes for full-duplex cooperative MIMO-NOMA systems

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Mobini, Zahra Mohammadi, Mohammadali Tsiftsis, Theodoros A. Ding, Zhiguo Tellambura, Chintha Faculty of Engineering Shahrekord University Shahrekord115 Iran School of Intelligent Systems Science and Engineering Jinan University Zhuhai519070 China Department of Computer Science and Telecommunications University of Thessaly Lamia35131 Greece School of Electrical and Electronic Engineering The University of Manchester Manchester United Kingdom Department of Electrical and Computer Engineering University of Alberta EdmontonABT6G 2V4 Canada

In this paper, we address the antenna selection (AS) problem in full-duplex (FD) cooperative non-orthogonal multiple access (NOMA) systems, where a multi-antenna FD relay bridges the connection between the multiantenna base station and NOMA far user. Specifically, two AS schemes, namely max-U1 and max-U2, are proposed to maximize the end-to-end signal-to-interference-plus-noise ratio at either or both near and far users, respectively. Moreover, a two-stage AS scheme, namely quality-of-service (QoS) provisioning scheme, is designed to realize a specific rate at the far user while improving the near user's rate. To enhance the performance of the QoS provisioning AS scheme, the idea of dynamic antenna clustering is applied at the relay to adaptively partition the relay's antennas into transmit and receive subsets. The proposed AS schemes' exact outage probability and achievable rate expressions are derived. To provide more insight, closed-form asymptotic outage probability expressions for the max-U1 and max-U2 AS schemes are obtained. Our results show that while the QoS provisioning AS scheme can deliver a near-optimal performance for static antenna setup at the relay, it provides up to 12% average sum rate gain over the optimum AS selection with fixed antenna setup. © 2022, CC BY.

关键词： Antennas

Explainable Artificial Intelligence to Detect Image Spam Using Convolutional Neural Network

学校读者我要写书评

暂无评论

Explainable Artificial Intelligence to Detect Image Spam Usi...

Cyber Resilience (ICCR), International Conference on

作者： Zhibo Zhang Ernesto Damiani Hussam Al Hamadi Chan Yeob Yeun Fatma Taher Department of Electrical Engineering and Computer Science C2PS Khalifa University Abu Dhabi United Arab Emirates Department of Computing and Applied Technology Zayed University Dubai United Arab Emirates

ISBN: (数字)9781665461221

ISBN: (纸本)9781665461238

Image spam threat detection has continually been a popular area of research with the internet's phenomenal expansion. This research presents an explainable framework for detecting spam images using Convolutional N eural Network (CNN) algorithms and Explainable Artificial Intelligence (XAI) algorithms. In this work, we use CNN model to classify image spam respectively whereas the post-hoc XAI methods including Local Interpretable Model Agnostic Explanation (LIME) and Shapley Additive Explanations (SHAP) were deployed to provide explanations for the decisions that the black-box CNN models made about spam image detection. We train and then evaluate the performance of the proposed approach on a 6636 image dataset including spam images and normal images collected from three different publicly available email corpora. The experimental results show that the proposed framework achieved satisfactory detection results in terms of different performance metrics whereas the model-independent XAI algorithms could provide explanations for the decisions of different models which could be utilized for comparison for the future study.

关键词： Measurement Additives Unsolicited e-mail Closed box Classification algorithms Convolutional neural networks Artificial intelligence

Systematizing Modeler Experience (MX) in Model-Driven engineering Success Stories

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Kalantari, Reyhaneh Oertel, Julian Exelmans, Joeri Rukmono, Satrio Adi Amaral, Vasco Tichy, Matthias Juhnke, Katharina Steghöfer, Jan-Philipp Abrahão, Silvia Department of Engineering Design and Teaching Innovation University of Ottawa 800 King Edward Avenue OttawaK1N6N5 Canada Department of Software Engineering University of Rostock Albert-Einstein-Straße Rostock18057 Germany Department of Computer Science University of Antwerp Middelheimlaan 1 Antwerp2020 Belgium Department of Mathematics and Computer Science Eindhoven University of Technology De Zaale Eindhoven5600 MB Netherlands Department of Computer Science NOVA School of Science & Technology Campus de Caparica Caparica2829-516 Portugal Institute of Software Engineering and Languages Ulm University James-Franck-Ring Ulm89081 Germany SMT-EMI6 Carl Zeiss SMT GmbH Rudolf-Eber-Straße 2 Oberkochen73447 Germany XITASO GmbH Austraße 35 Augsburg86153 Germany Instituto Universitario Mixto de Tecnología Informática Universitat Politècnica de València Camino de Vera s/n Valencia46022 Spain

Modeling is often associated with complex and heavy tooling, leading to a negative perception among practitioners. However, alternative paradigms, such as everything-as-code or low-code, are gaining acceptance due to their perceived ease of use. This paper explores the dichotomy between these perceptions through the lens of "modeler experience" (MX). MX includes factors such as user experience, motivation, integration, collaboration & versioning and language complexity. We examine the relationships between these factors and their impact on different modeling usage scenarios. Our findings highlight the importance of considering MX when understanding how developers interact with modeling tools and the complexities of modeling and associated tooling. © 2024, CC BY.

关键词：

Application-Dependent Support of ITS-G5 Car-to-Car Communication by Selected Usage of LTE-V2X Resources

学校读者我要写书评

暂无评论

Application-Dependent Support of ITS-G5 Car-to-Car Communica...

AmE 2021 - Automotive meets Electronics;12th GMM-Symposium

作者： Jochen Stellwagen Matthias Deegener Michael Kuhn Frankfurt University of Applied Sciences Facility 2: Computer Science and Engineering Darmstadt University of Applied Sciences Department of Electrical Engineering

ISBN: (纸本)9783800754878

Car-to-car communication is besides autonomous driving and e-mobility one of today’s most important topics in automotive research. Currently, two transmission technologies are available for safety-critical applications in vehicle-to-vehicle communication: WIFI-based ITS-G5 and cellular-based LTE-V2X [1]. In this article ITS-G5 is used as the primary transmission technology to support safety-critical applications within a wide range of traffic situations by exchanging information between individual vehicles. The resources of the LTE-V2X communication will be used to support the WIFI-based communication in case a reliable transmission via ITS-G5 is not or only partially possible. In a simulation environment several test series were developed to investigate different traffic scenarios, use cases and equipment rates for both transmission technologies. This will show how often and in which situations the resources of LTE-V2X are used.

关键词： Communication Research transmission Simulated Environment Vehicle to vehicle communications Safety critical Traffic

SINDy-RL: Interpretable and Efficient Model-Based Reinforcement Learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zolman, Nicholas Fasel, Urban Nathan Kutz, J. Brunton, Steven L. Department of Mechanical Engineering University of Washington SeattleWA98195 United States Department of Aeronautics Imperial College LondonSW7 2AZ United Kingdom Department of Applied Mathematics University of Washington SeattleWA98195 United States Department of Electrical and Computer Engineering University of Washington SeattleWA98195 United States Data Science and Artificial Intelligence Department The Aerospace Corporation El SegundoCA90245 United States

Deep reinforcement learning (DRL) has shown significant promise for uncovering sophisticated control policies that interact in environments with complicated dynamics, such as stabilizing the magnetohydrodynamics of a tokamak fusion reactor or minimizing the drag force exerted on an object in a fluid flow. However, these algorithms require an abundance of training examples and may become prohibitively expensive for many applications. In addition, the reliance on deep neural networks often results in an uninterpretable, black-box policy that may be too computationally expensive to use with certain embedded systems. Recent advances in sparse dictionary learning, such as the sparse identification of nonlinear dynamics (SINDy), have shown promise for creating efficient and interpretable data-driven models in the low-data regime. In this work we introduce SINDy-RL, a unifying framework for combining SINDy and DRL to create efficient, interpretable, and trustworthy representations of the dynamics model, reward function, and control policy. We demonstrate the effectiveness of our approaches on benchmark control environments and challenging fluids problems. SINDy-RL achieves comparable performance to state-of-the-art DRL algorithms using significantly fewer interactions in the environment and results in an interpretable control policy orders of magnitude smaller than a deep neural network policy. © 2024, CC BY.

关键词： Reinforcement learning