检索结果-内蒙古大学图书馆

Dense Pixel-to-Pixel Harmonization via Continuous Image Representation

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Chen, Jianqi Zhang, Yilan Zou, Zhengxia Chen, Keyan Shi, Zhenwei Image Processing Center School of Astronautics Beihang University Beijing100191 China State Key Laboratory of Virtual Reality Technology and Systems Beihang University Beijing100191 China Shanghai Artificial Intelligence Laboratory Shanghai200232 China Department of Guidance Navigation and Control School of Astronautics Beihang University Beijing100191 China

High-resolution (HR) image harmonization is of great significance in real-world applications such as image synthesis and image editing. However, due to the high memory costs, existing dense pixel-to-pixel harmonization methods are mainly focusing on processing low-resolution (LR) images. Some recent works resort to combining with color-to-color transformations but are either limited to certain resolutions or heavily depend on hand-crafted image filters. In this work, we explore leveraging the implicit neural representation (INR) and propose a novel image Harmonization method based on Implicit neural Networks (HINet), which to the best of our knowledge, is the first dense pixel-to-pixel method applicable to HR images without any handcrafted filter design. Inspired by the Retinex theory, we decouple the MLPs into two parts to respectively capture the content and environment of composite images. A Low-Resolution Image Prior (LRIP) network is designed to alleviate the Boundary Inconsistency problem, and we also propose new designs for the training and inference process. Extensive experiments have demonstrated the effectiveness of our method compared with state-of-the-art methods. Furthermore, some interesting and practical applications of the proposed method are explored. Our code is available at https://***/WindVChen/INR-Harmonization. Copyright © 2023, The Authors. All rights reserved.

关键词： Pixels

Non-Euclidean Monotone Operator Theory with Applications to Recurrent Neural Networks

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Davydov, Alexander Jafarpour, Saber Proskurnikov, Anton V. Bullo, Francesco The Center for Control Dynamical Systems and Computation University of California Santa Barbara93106-5070 United States The School of Electrical and Computer Engineering Georgia Institute of Technology United States The Department of Electronics and Telecommunications Politecnico di Torino Turin Italy

We provide a novel transcription of monotone operator theory to the non-Euclidean finite-dimensional spaces l1 and l∞. We first establish properties of mappings which are monotone with respect to the non-Euclidean norms l1 or l∞. In analogy with their Euclidean counterparts, mappings which are monotone with respect to a non-Euclidean norm are amenable to numerous algorithms for computing their zeros. We demonstrate that several classic iterative methods for computing zeros of monotone operators are directly applicable in the non-Euclidean framework. We present a case-study in the equilibrium computation of recurrent neural networks and demonstrate that casting the computation as a suitable operator splitting problem improves convergence rates. Copyright © 2022, The Authors. All rights reserved.

关键词： Recurrent neural networks

Secure State Estimation against Sparse Attacks on a Time-varying Set of Sensors

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Li, Zishuo Niazi, Muhammad Umar B. Liu, Changxin Mo, Yilin Johansson, Karl H. Department of Automation Tsinghua University Beijing100084 China Division of Decision and Control Systems Digital Futures School of Electrical Engineering and Computer Science KTH Royal Institute of Technology StockholmSE-100 44 Sweden Laboratory for Information and Decision Systems Massachusetts Institute of Technology 77 Massachusetts Avenue CambridgeMA02139 United States

This paper studies the problem of secure state estimation of a linear time-invariant (LTI) system with bounded noise in the presence of sparse attacks on an unknown, time-varying set of sensors. In other words, at each time, the attacker has the freedom to choose an arbitrary set of no more that p sensors and manipulate their measurements without restraint. To this end, we propose a secure state estimation scheme and guarantee a bounded estimation error subject to 2p-sparse observability and a mild, technical assumption that the system matrix has no degenerate eigenvalues. The proposed scheme comprises a design of decentralized observer for each sensor based on the local observable subspace decomposition. At each time step, the local estimates of sensors are fused by a median operator to obtain a secure estimation, which is then followed by a local detection-and-resetting process of the decentralized observers. The estimation error is shown to be upper-bounded by a constant which is determined only by the system parameters and noise magnitudes. Moreover, we design the detector threshold to ensure that the benign sensors never trigger the detector. The efficacy of the proposed algorithm is demonstrated by its application on a benchmark example of IEEE 14-bus system. © 2022, CC BY-NC-ND.

关键词： Fault detection

Active wheel speed control to avoid lifting the swingarms in rocker-bogie suspension 3

学校读者我要写书评

暂无评论

Active wheel speed control to avoid lifting the swingarms in...

3rd International Young Engineers Forum, YEF-ECE 2019

作者： Tomaszuk, P. Lukowska, A. Recko, M. Dzierzek, K. Straszynski, P. Faculty of Mechanical Engineering Bialystok University of Technology Department of Automatic Control and Robotics Bialystok Poland

ISBN: (纸本)9781538692820

Rocker-bogie is kind of suspension system where wheels from each side of the vehicle are connected with swingarm on the rotary axle. Both levers are connected with a differential mechanism. If the vehicle has not had turnable wheels and turns only using the velocity or heading difference of each side or the object, then increased traction on the most loaded wheels causes lifting of the swingarms in opposing directions. This situation is undesirable because it leads to the high inclination of the vehicle's body and threatens to fall over. This article shows the way of active adjusting each wheel's speed to avoid swingarms lifting. Presented algorithm increases or decreases the angular velocity of selected wheels depending on the robot's rotation speed, acceleration, and traction. Parameters are calculated from the velocity of swingarm lift angle changes compared to the vehicle's relative rotation speed. Theoretical description presents the geometry of vehicle suspension, used sensors, and wheels speed control circuitry. The algorithm is implemented on an existing vehicle, and experimental part contains a comparison of behavior with and without active control on different surfaces. © 2019 IEEE.

关键词： Automobile suspensions

Deep Learning Based Algorithm for automatic Scoliosis Angle Measurement 3

学校读者我要写书评

暂无评论

Deep Learning Based Algorithm for Automatic Scoliosis Angle ...

3rd International Conference on Computer Applications and Information Security, ICCAIS 2020

作者： Alharbi, Roa H. Alshaye, Meshal B. Alkanhal, Maryam M. Alharbi, Najla M. Alzahrani, Mosa A. Alrehaili, Osama A. King Abdulaziz City for Science and Technology National Center for Communication Systems and Command and Control Technology Riyadh Saudi Arabia King Saud University Medical City Department of Orthopedic Riyadh Saudi Arabia

ISBN: (纸本)9781728142128

Scoliosis is a common back disease which identifies with an irregular spinal condition. In this case, the spine has a side curvature with an angle. Practically, the standard angle estimation method is done by measuring the Cobb angle for the curvature. Cobb angle is the angle between two drawn lines, upper-end line and lower-end line of the curve. However, manual measurement needs time and effort. In this paper, we proposed an automatic measurement algorithm with machine learning. Initially, X-Rays images are processed utilizing CLAHE method. Then, deep convolutional neural networks (CNN) are applied to detect vertebrae in each X-Ray image. At last, the Cobb angle is measured through a novel algorithm using trigonometry. The proposed method is evaluated on X-Rays dataset from King Saud University (KSU), and it detects each vertebra in those images. In addition, Cobb angle measurements are compared with experts' manual measurements. Our method achieves the estimation of Cobb angles with high accuracy, showing its great potential in clinical use. © 2020 IEEE.

关键词： Convolutional neural networks

AI Graphical Programming Learning Platform for Children

学校读者我要写书评

暂无评论

AI Graphical Programming Learning Platform for Children

2021 China Automation Congress, CAC 2021

作者： Zhao, Hongxia Liu, Xiwei Gong, Xiaoyan Li, Qiang Jing, Sifeng Xue, Yaofeng The State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China Institute of Smart Education Systems Qingdao Academy of Intelligent Industries Qingdao China University of Chinese Academy of Sciences Beijing China Shanghai Engineering Research Center of Digital Education Equipment Department of Education Information Technology East China Normal University Shanghai China

ISBN: (纸本)9781665426473

In this era of rapid development of artificial intelligence, the truth of competition in the future is the talent. Countries all over the world attach great importance to the programming and artificial intelligence education of children. In the future, the theoretical development and practical application of this emerging field will usher in a period of rapid development. This paper presents an artificial intelligence graphical programming learning platform based on Jetson Nano, which combines artificial intelligence education with modular software programming, provides students with a more open, intelligent, and interesting programming environment through the cross-platform development mode of "cloud service + intelligent hardware", and adopts visual programming to reduce the difficulty of artificial intelligence programming. The platform can fully tap learning resources, reduce education costs, and make artificial intelligence education more intuitive and humanized. © 2021 IEEE

关键词： Computer graphics

Concept, Principle, and Modeling of Driving Risk Entropy Based on Human-Vehicle-Road Coupling Model for Autonomous Vehicle

学校读者我要写书评

暂无评论

SSRN

SSRN 2024年

作者： Gao, Hongbo Yang, Hanqing Zhu, Juping Su, Huiping Wang, Ning Xie, Liping Li, Keqiang Department of Automation University of Science and Technology of China Hefei230026 China Institute of Advanced Technology University of Science and Technology of China Hefei230088 China School of Electrical and Electronic Engineering Nanyang Technological University 639798 Singapore Chongqing College of Mobile Communication Chongqing400044 China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education School of Automation Southeast University Nanjing210096 China State Key Laboratory of Automotive Safety and Energy Tsinghua University Beijing100084 China

Driving risk entropy, based on entropy law, is an innovative concept proposed for intelligent driving systems. The concept deals with the driving risks caused by the human-vehicle-road system from the driving information obtained;however, driving risks are difficult to assess due to many influencing factors and great uncertainty. The risk entropy model of vehicles is visualized by introducing the concept of basic isotropic risk elliptic line (BIREL). Further incorporating the concept of information entropy, risk entropy models for vehicle interaction and vulnerable participants are established, and the risk entropy model for roads is developed using raster coordinates. Ultimately, a three-dimensional driving risk entropy model is constructed based on the human-vehicle-road coupling model, with appropriate weights assigned to each parameter through the DDPG algorithm. Experimental results demonstrate the effectiveness of the driving risk entropy model. © 2024, The Authors. All rights reserved.

关键词： Autonomous vehicles

Implicit Ray-Transformers for Multi-view Remote Sensing Image Segmentation

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Qi, Zipeng Chen, Hao Liu, Chenyang Shi, Zhenwei Zou, Zhengxia The Image Processing Center School of Astronautics The Beijing Key Laboratory of Digital Media The State Key Laboratory of Virtual Reality Technology and Systems Beihang University Beijing100191 China The Shanghai Artificial Intelligence Laboratory Shanghai200232 China The Department of Guidance Navigation and Control School of Astronautics Beihang University Beijing100191 China

The mainstream CNN-based remote sensing (RS) image semantic segmentation approaches typically rely on massive labeled training data. Such a paradigm struggles with the problem of RS multi-view scene segmentation with limited labeled views due to the lack of considering 3D information within the scene. In this paper, we propose "Implicit Ray-Transformer (IRT)" based on Implicit Neural Representation (INR), for RS scene semantic segmentation with sparse labels (such as 4-6 labels per 100 images). We explore a new way of introducing multi-view 3D structure priors to the task for accurate and view-consistent semantic segmentation. The proposed method includes a two-stage learning process. In the first stage, we optimize a neural field to encode the color and 3D structure of the remote sensing scene based on multi-view images. In the second stage, we design a Ray Transformer to leverage the relations between the neural field 3D features and 2D texture features for learning better semantic representations. Different from previous methods that only consider 3D prior or 2D features, we incorporate additional 2D texture information and 3D prior by broadcasting CNN features to different point features along the sampled ray. To verify the effectiveness of the proposed method, we construct a challenging dataset containing six synthetic sub-datasets collected from the Carla platform and three real sub-datasets from Google Maps. Experiments show that the proposed method outperforms the CNN-based methods and the state-of-the-art INR-based segmentation methods in quantitative and qualitative metrics. Ablation study shows that under limited label conditions, the combination of the 3D structure prior and 2D texture can significantly improve the performance and effectively complete missing semantic information in novel views. Experiments also demonstrate the proposed method could yield geometry-consistent segmentation results against illumination changes and viewpoint changes. Our dat

关键词： Remote sensing

Dragonfly Optimization-based MPPT Algorithm for Standalone PV System under Partial Shading

学校读者我要写书评

暂无评论

Dragonfly Optimization-based MPPT Algorithm for Standalone P...

2021 IEEE International Conference on Emergency Science and Information technology, ICESIT 2021

作者： Lodhi, Ehtisham Yang, Pu Wang, Lina Khan, M. Adil Lodhi, Zeeshan Javed, Usman Saleem, Qasim School of Automation Science and Electrical Engineering Beihang University Beijing100191 China SKL for Management and Control of Complex Systems Institute OfAutomation Chinese Academy of Sciences Beijing100190 China Chang'an University Department of Computer and Technology Xi'an710062 China COMSATS University Department of Electrical and Computer Engineering Islamabad22060 Pakistan Lahore54000 Pakistan

ISBN: (纸本)9781665435314

Currently, Photovoltaic (PV) technology has higher demand in the market and gaining momentum day by day. However, partial shading is the most significant concern as it reduces the efficiency of standalone PV systems. Partially shading is typically induced by non-uniform sunlight and uncertain temperature fluctuations. The maximum power point tracking (MPPT) algorithms play a major role in determining the overall system's efficiency. The PV system creates several peaks in its power voltage (P-V) characteristic curve under partial shadowing. To make the PV system run under partial shading at global maximum power point (GMPP), the dragonfly optimization (DFO) algorithm is put forward in this research article. The proposed DFO technique and other commonly used MPPT techniques, such as Particle swarm optimization (PSO), Perturb and Observe (PO), is investigated through theoretical analysis and simulation using Matlab/Simulink to authenticate the overall performance of the DFO approach under partial shading. Furthermore, the comparison demonstrates that the suggested DFO strategy provides very high reliability, efficiency, robustness, rapid power tracking and lower response time to trace the GMPP than PSO and PO techniques. Hence, the DFO technique's overall performance under partial shading is appropriate and satisfying. © 2021 IEEE.

关键词： Simulink