检索结果-内蒙古大学图书馆

EGRA-NeRF: Edge-Guided Ray Allocation for Neural Radiance Fields

学校读者我要写书评

暂无评论

Image and Vision Computing 2023年 134卷

作者： Gai, Zhenbiao Liu, Zhenyang Tan, Min Ding, Jiajun Yu, Jun Tong, Mingzhao Yuan, Junqing School of Computer Science Hangzhou Dianzi University Zhejiang Hangzhou310018 China Department of Applied Mathematics School of Science Zhejiang University of Technology Zhejiang Hangzhou310023 China

Recently, Neural Radiance Fields (NeRF) has demonstrated great potential in synthesizing novel views for realistic video generation. However, renderings from NeRF appear excessively blurred and contain aliasing artifacts in some textures or edges. To alleviate this problem, Edge-Guided Ray Allocation (EGRA-NeRF) module is proposed in this paper. Such method proposes a novel ray allocation strategy to focus more rays on the textures and edges of the scene during the training stage. Specially, the Canny edge detector is introduced to generate the ray guidance map. This guidance map can lead the model to find areas with more textures. To focus more rays on edges and textures, a method is explored to re-adjust the ray guidance map and re-allocate rays dynamically. Compared with the ray allocation in most existing NeRF-based methods, EGRA-NeRF focuses rays more on textures and edges, which can achieve the refinement of edges and texture details. Experiments show that NeRF-based algorithms (NeRF and Instant-NGP) with our EGRA module can achieve more realistic video generation, while maintaining almost the same computation time. © 2023

关键词： Textures

Forecasting of Fires Hotspot in Tropical Region Using LSTM Algorithm Based on Satellite Data

学校读者我要写书评

暂无评论

Forecasting of Fires Hotspot in Tropical Region Using LSTM A...

2022 IEEE Region 10 Symposium, TENSYMP 2022

作者： Kadir, Evizal Abdul Kung, Hsiang Tsung Rosa, Sri Listia Sabot, Andrew Othman, Mahmod Ting, Mark Universitas Islam Riau Department of Informatics Engineering Pekanbaru28284 Indonesia Harvard University Department of Computer Science CambridgeMA02134 United States Universiti Teknologi Petronas Department of Applied Mathematics Perak86400 Malaysia

ISBN: (数字)9781665466585

ISBN: (纸本)9781665466585

Raising of temperature due to global warming impacted to the number of fires hotspot globally, in tropical region some of the places a high risk such as wildfire. Fire in Indonesia is one of the big disasters because of forestry region and peat land that highly risk to wildfire especially in dry season. This research aims to do a forecasting number of fires hotspot in future time based on training data collected in previous years. Long Short-Term Memory (LSTM) algorithm applied in this forecasting, the advantages of LSTM model to forecast a timeseries data make the high prediction accuracy. Fires hotspot data was collected since year 2010 to 2021 with total number of datasets more than 700,000 hotspot and forecast for the future year 2022. Collected datasets analyzed use LSTM model divided into two classification which are training and test data with 70% and 30% respectively. Results shows a similar pattern of forecasting fires hotspot in 2022 compared into previous last 2 years which 2021 and 2020. Furthermore, to proof proposed algorithm is working fine then a forecast number of fire hotspot for year 2021 have been done which compared actual and forecasting data and percentage of error 4.56%. LSTM Algorithm is one of the models suitable to use in data forecasting in high volume time series data. © 2022 IEEE.

关键词： Fires

Convergence of Equilibrium Measures under K-regular Polynomial Sequences and their Derivatives

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Henriksen, Christian Petersen, Carsten Lunde Uhre, Eva Department of Applied Mathematics and Computer Science Technical University of Denmark Denmark Department of Mathematical Sciences Copenhagen University Denmark Department of Science and Environment Roskilde University Denmark

Let K ⊂ ℂ be non-polar, compact and polynomially convex. We study the limits of equilibrium measures on preimages of compact sets, under K-regular sequences of polynomials, that center on K and under the sequences of derivatives of all orders of such sequences. We show that under mild assumptions such limits always exist and equal the equilibrium measure on K. From this we derive convergence of the equilibrium distributions on the Julia sets of the sequence of polynomials and their derivatives of all *** Codes 31A05 (Primary) 37F10, 42C05 (Secondary) Copyright © 2023, The Authors. All rights reserved.

关键词： Polynomials

CFD-DRIVEN TOPOLOGY OPTIMIZATION FOR PERSONALIZED INTRACRANIAL ANEURYSM IMPLANT DESIGN

学校读者我要写书评

暂无评论

CFD-DRIVEN TOPOLOGY OPTIMIZATION FOR PERSONALIZED INTRACRANI...

ASME 2024 International Design Engineering Technical Conferences and computers and Information in Engineering Conference, IDETC-CIE 2024

作者： Lei, Weijun Sadasivan, Chander Gu, Xianfeng David Chen, Shikui Department of Mechanical Engineering State University of New York at Stony Brook Stony BrookNY11794 United States Department of Neurological Surgery Stony Brook University Medical Center Stony BrookNY11794 United States Department of Computer Science Department of Applied Mathematics and Statistics State University of New York at Stony Brook Stony BrookNY11794 United States

ISBN: (纸本)9780791888377

Intracranial aneurysm rupture causes life-threatening subarachnoid hemorrhage. Current endovascular devices like coils, flow diverters, and intravascular implants aim to thrombose the aneurysm but have limitations and varying success rates depending on aneurysm characteristics. We propose a new computational framework integrating CFD and topology optimization to design personalized aneurysm implants. The optimization problem aims to reduce blood flow velocity within the aneurysm while ensuring adequate structural integrity of the implant. The fluid dynamics are governed by the Navier-Stokes equations, while the solid mechanics are described by the linear elasticity equations. A Darcy-Brinkman model is employed to simulate flow through the porous implant in the fluid domain, while the Solid Isotropic Material with Penalization (SIMP) method is used to interpolate between solid and void regions in the structural domain during topology optimization. The objective combines fluid energy dissipation ratio and solid strain energy with spatially varying weights. Global and local volume constraints generate personalized implants with porosity and flow-diverting architectures. The approach is demonstrated on patient-specific aneurysm geometries from rotational angiography. This CFD-driven topology optimization method enables personalized aneurysm implant design to potentially improve occlusion rates and reduce complications compared to current devices. Further studies will validate the optimized designs and investigate their efficacy in vitro and in vivo. Copyright © 2024 by ASME.

关键词： Navier Stokes equations

G-NM: A Group of Numerical Time Series Prediction Models

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Yun, Juyoung Department of Computer Science Stony Brook University United States Department of Applied Mathematics and Statistics Stony Brook University United States

In this study, we focus on the development and implementation of a comprehensive ensemble of numerical time series forecasting models, collectively referred to as the Group of Numerical Time Series Prediction Model (G-NM). This inclusive set comprises traditional models such as Autoregressive Integrated Moving Average (ARIMA), Holt-Winters' method, and Support Vector Regression (SVR), in addition to modern neural network models including Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM). G-NM is explicitly constructed to augment our predictive capabilities related to patterns and trends inherent in complex natural phenomena. By utilizing time series data relevant to these events, G-NM facilitates the prediction of such phenomena over extended periods. The primary objective of this research is to both advance our understanding of such occurrences and to significantly enhance the accuracy of our forecasts. G-NM encapsulates both linear and non-linear dependencies, seasonalities, and trends present in time series data. Each of these models contributes distinct strengths, from ARIMA's resilience in handling linear trends and seasonality, SVR's proficiency in capturing non-linear patterns, to LSTM's adaptability in modeling various components of time series data. Through the exploitation of the G-NM potential, we strive to advance the state-of-the-art in large-scale time series forecasting models. We anticipate that this research will represent a significant stepping stone in our ongoing endeavor to comprehend and forecast the complex events that constitute the natural world. © 2023, CC0.

关键词： Long short-term memory

What Is So Deep About Deepfakes? A Multi-Disciplinary Thematic Analysis of Academic Narratives About Deepfake Technology

学校读者我要写书评

暂无评论

IEEE Transactions on Technology and Society

IEEE Transactions on Technology and Society 2025年第1期6卷 64-79页

作者： Twomey, John Ching, Didier Aylett, Matthew Peter Quayle, Michael Linehan, Conor Murphy, Gillian University College Cork School of Applied Psychology CorkT12 CY82 Ireland Lero—The Science Foundation Ireland Research Centre for Software LimerickV94 NYD3 Ireland Heriot Watt University Mathematics and Computer Science Department EdinburghEH14 4AS United Kingdom CereProc Ltd. EdinburghEH3 9DR United Kingdom University of Limerick Centre for Social Issues Research Department of Psychology LimerickV94 T9PX Ireland University of KwaZulu-Natal School of Applied Human Sciences Department of Psychology Pietermaritzburg3201 South Africa

Deepfakes are a form of synthetic media that uses deep-learning technology to create fake images, video, and audio. The emergence of this technology has inspired much commentary and speculation from academics across a range of disciplines, who have contributed expert opinions regarding the implications of deepfake proliferation on fields such as law, politics, and entertainment. A systematic scoping review was carried out to identify, assemble, and critically analyze those academic narratives. The aim is to build on and critique previous attempts at defining the technology and categorizing the harms and benefits of deepfake technology. A range of databases were searched for relevant articles from 2017 to 2023, resulting in a large multi-disciplinary dataset of 102 papers, 181,659 words long, which were analyzed qualitatively through thematic analysis. Implications for future research include questioning the lack of research evidence for the supposed positives of deepfakes, recognizing the role that identity plays in deepfake technology, challenging the perceived accessibility/ believability of deepfakes, and proposing a more nuanced approach to the dichotomous "positive and negatives" of deepfakes. Furthermore, we show how definitional issues around what a deepfake is versus other forms of fake media feeds confusion around the novelty and impacts of deepfakes. © 2024 The Authors.

关键词： Deepfakes Reviews Generative AI Software Psychology Social networking (online) Organizations Lenses Law Internet

Stochastic Gradient Sampling for Enhancing Neural Networks Training

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Yun, Juyoung Stony Brook University Department of Computer Science United States Stony Brook University Department of Applied Mathematics and Statistics United States

In this paper, we introduce StochGradAdam, a novel optimizer designed as an extension of the Adam algorithm, incorporating stochastic gradient sampling techniques to improve computational efficiency while maintaining robust performance. StochGradAdam optimizes by selectively sampling a subset of gradients during training, reducing the computational cost while preserving the advantages of adaptive learning rates and bias corrections found in Adam. Our experimental results, applied to image classification and segmentation tasks, demonstrate that StochGradAdam can achieve comparable or superior performance to Adam, even when using fewer gradient updates per iteration. By focusing on key gradient updates, StochGradAdam offers stable convergence and enhanced exploration of the loss landscape, while mitigating the impact of noisy gradients. The results suggest that this approach is particularly effective for large-scale models and datasets, providing a promising alternative to traditional optimization techniques for deep learning applications. © 2023, CC0.

关键词： Contrastive Learning

IIoT-based Predictive Maintenance for Oil and Gas Industry 22

学校读者我要写书评

暂无评论

IIoT-based Predictive Maintenance for Oil and Gas Industry

Proceedings of the 2022 6th International Conference on Electronic Information Technology and computer Engineering

作者： Zhiyang Jia Jihe Wang Cheng Deng Department of Computer Science China University of Petroleum-Beijing at Karamay China Department of Mathematics West Yunnan University of Applied Sciences China

ISBN: (纸本)9781450397148

The Industrial Internet of Things (IIoT) promises to provide an expanded awareness of field assets and equipment, access to data from across locations, and actionable insights for maximizing operational performance and safety of the oil and gas industry. Using automation and machine learning, with the application of predictive maintenance, efficiencies can be boosted and problems can be mitigated sooner and more effectively. The proposed system is mainly based on the data collection, processing, analysis, and modeling of an enormous number of historical and real-time data generated during the operation of the equipment on the edge side. The data-driven predictive maintenance used machine learning models and deep learning models to predict the remaining useful life (RUL). Bi-LSTM based prediction model has been trained on the cloud, and deployed onto the edge devices. The predictive maintenance process includes data acquisition, data processing, training of machine learning model, equipment health assessment, remaining useful life prediction, strategy formulation, and strategy execution. The predictive maintenance solution driven by the IIoT helps oil and gas companies make predictions before equipment failures have a significant impact on their company's safety level and profits to improve asset reliability and promote cost savings.

关键词： Oil and Gas Industry IIot Long short-term memory Predictive maintenance

Value Distributions of Derivatives of K-regular Polynomial Families

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Henriksen, Christian Petersen, Carsten Lunde Uhre, Eva Department of Applied Mathematics and Computer Science Technical University of Denmark Denmark Department of Mathematical Sciences University of Copenhagen Denmark Department of Science and Environment Roskilde University Denmark

Let Ω ∈ C be a domain such that K := C\Ω is compact and non-polar. Let (qk)k>0 be a sequence of polynomials with nk, the degree of qk satisfying nk → ∞, and let (qk(m))k denote the sequence of m-th derivatives. We provide conditions, which ensure that the preimages (qk(m))−1({a}) uniformly equidistribute on ∂Ω, as k → ∞, for every a ∈ C and every m = 0, 1, . . .MSC Codes 42C05 (Primary) 37F10, 31A15 (Secondary) Copyright © 2023, The Authors. All rights reserved.

关键词： Green's function