检索结果-内蒙古大学图书馆

Neural human video rendering by learning dynamic textures and rendering-to-video translation

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Liu, Lingjie Xu, Weipeng Habermann, Marc Zollhöfer, Michael Bernard, Florian Kim, Hyeongwoo Wang, Wenping Theobalt, Christian The Department of Computer Science The University of Hong Kong Hong Kong The Graphics Vision and Video Group Max Planck Institute for Informatics Saarbrücken66123 Germany The Department of Computer Science Computer Graphics Laboratory Stanford University CA94305 United States

Synthesizing realistic videos of humans using neural networks has been a popular alternative to the conventional graphics-based rendering pipeline due to its high efficiency. Existing works typically formulate this as an image-to-image translation problem in 2D screen space, which leads to artifacts such as over-smoothing, missing body parts, and temporal instability of fine-scale detail, such as pose-dependent wrinkles in the clothing. In this paper, we propose a novel human video synthesis method that approaches these limiting factors by explicitly disentangling the learning of time-coherent fine-scale details from the embedding of the human in 2D screen space. More specifically, our method relies on the combination of two convolutional neural networks (CNNs). Given the pose information, the first CNN predicts a dynamic texture map that contains time-coherent high-frequency details, and the second CNN conditions the generation of the final video on the temporally coherent output of the first CNN. We demonstrate several applications of our approach, such as human reenactment and novel view synthesis from monocular video, where we show significant improvement over the state of the art both qualitatively and quantitatively. Copyright © 2020, The Authors. All rights reserved.

关键词： Convolutional neural networks

Beyond identity: what information is stored in biometric face templates?

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Terhörst, Philipp Fährmann, Daniel Damer, Naser Kirchbuchner, Florian Kuijper, Arjan Fraunhofer Institute for Computer Graphics Research IGD Darmstadt64283 Germany Department of Computer Science Technical University of Darmstadt Darmstadt64289 Germany

Deeply-learned face representations enable the success of current face recognition systems. Despite the ability of these representations to encode the identity of an individual, recent works have shown that more information is stored within, such as demographics, image characteristics, and social traits. This threatens the user’s privacy, since for many applications these templates are expected to be solely used for recognition purposes. Knowing the encoded information in face templates helps to develop bias-mitigating and privacy-preserving face recognition technologies. This work aims to support the development of these two branches by analysing face templates regarding 113 attributes. Experiments were conducted on two publicly available face embeddings. For evaluating the predictability of the attributes, we trained a massive attribute classifier that is additionally able to accurately state its prediction confidence. This allows us to make more sophisticated statements about the attribute predictability. The results demonstrate that up to 74 attributes can be accurately predicted from face templates. Especially non-permanent attributes, such as age, hairstyles, haircolors, beards, and various accessories, found to be easily-predictable. Since face recognition systems aim to be robust against these variations, future research might build on this work to develop more understandable privacy preserving solutions and build robust and fair face templates. Copyright © 2020, The Authors. All rights reserved.

关键词： Face recognition

Reasoning about Human-Friendly Strategies in Repeated Keyword Auctions 22

学校读者我要写书评

暂无评论

Reasoning about Human-Friendly Strategies in Repeated Keywor...

Proceedings of the 21st International Conference on Autonomous Agents and Multiagent Systems

作者： Francesco Belardinelli Wojtek Jamroga Vadim Malvone Munyque Mittelmann Aniello Murano Laurent Perrussel Université d'Evry Évry France University of Luxembourg & Institute of Computer Science Polish Academy of Sciences Luxembourg Luxembourg Télécom Paris Paris France Université de Toulouse - IRIT Toulouse France University of Naples Federico II Naples Italy

ISBN: (纸本)9781450392136

In online advertising, search engines sell ad placements for keywords continuously through auctions. This problem can be seen as an infinitely repeated game since the auction is executed whenever a user performs a query with the keyword. As advertisers may frequently change their bids, the game will have a large set of equilibria with potentially complex strategies. In this paper, we propose the use of natural strategies for reasoning in such setting as they are processable by artificial agents with limited memory and/or computational power as well as understandable by human users. To reach this goal, we introduce a quantitative version of Strategy Logic with natural strategies in the setting of imperfect information. In a first step, we show how to model strategies for repeated keyword auctions and take advantage of the model for proving properties evaluating this game. In a second step, we study the logic in relation to the distinguishing power, expressivity, and model-checking complexity for strategies with and without recall.

关键词： auctions mechanism design strategic reasoning

A comprehensive review of dwell time optimization methods in computer-controlled optical surfacing

学校读者我要写书评

暂无评论

Light(Advanced Manufacturing) 2024年第3期5卷 142-171页

作者： Tianyi Wang Xiaolong Ke Lei Huang Qingqing Cui Zili Zhang Chunjin Wang Hyukmo Kang Weslin Pullen Heejoo Choi Daewook Kim Vipender Negi Qian Kemao Yi Zhu Stefano Giorgio Philip Boccabella Nathalie Bouet Corey Austin Mourad Idir National Synchrotron Light Source II(NSLS-II) Brookhaven National LaboratoryP.O.Box 5000UptonNY 11973USA School of Mechanical and Automotive Engineering Xiamen University of TechnologyXiamen 361024China State Key Laboratory of Ultra-precision Machining Technology Department of Industrial Systems and EngineeringThe Hong Kong Polytechnic UniversityHung HomKowloonHong KongChina James C.Wyant College of Optical Sciences University of Arizona1630 E.University Blvd.P.O.Box 210094TucsonAZ 85721-0094USA Large Binocular Telescope Observatory University of ArizonaTucsonAZ 85721USA Department of Astronomy and Steward Observatory University of Arizona933 N.Cherry Ave.TucsonAZ85721USA Indian Institute of Technology Delhi Hauz KhasNew Delhi 110016India School of Computer Science and Engineering Nanyang Technological University50 Nanyang AvenueSingapore 639798

Dwell time plays a vital role in determining the accuracy and convergence of the computer-controlled optical surfacing ***,optimizing dwell time presents a challenge due to its ill-posed nature,resulting in non-unique *** address this issue,several well-known methods have emerged,including the iterative,Bayesian,Fourier transform,and matrix-form *** their independent development,these methods share common objectives,such as minimizing residual errors,ensuring dwell time's positivity and smoothness,minimizing total processing time,and enabling flexible dwell *** paper aims to comprehensively review the existing dwell time optimization methods,explore their interrelationships,provide insights for their effective implementations,evaluate their performances,and ultimately propose a unified dwell time optimization methodology.

关键词： Dwell time optimization computer-controlled optical surfacing Finishing Polishing Optical fabrication

Automatic diagnosis of disc herniation based on DenseNet fusion model 8

学校读者我要写书评

暂无评论

Automatic diagnosis of disc herniation based on DenseNet fus...

8th International Conference on Digital Home, ICDH 2020

作者： Wang, Zimin Qin, Junyan Huang, Jinrong Wang, Yumeng Li, Ji Guilin University of Electronic Technology Guangxi Colleges and Universities Key Lab. of Intelligent Processing of Computer Image and Graphics Guilin541004 China Guilin Huigu Institute of Artificial Intelligence Industrial Technology Guilin541004 China Guilin University of Electronic Technology School of Computer Science and Information Security Guilin541004 China

ISBN: (纸本)9781728192345

Spine MRI images generally have the characteristics of low contrast and much noise. Because the variable shape of the spine edge, the traditional spine image segmentation method requires a lot of preprocessing and can not perform image segmentation automatically, which is inefficient and cannot compliant the clinical work requirements. So as to solve the above problems, this paper use deep learning algorithms to identify the spine images, and proposed a disc herniation recognition method for spinal MRI images which based on DenseNet fusion model. Use DenseNet and Inception-Resnet-V2 model fusion to get more image features and improve the image recognition accuracy. Experiments prove that this method can be used to identify the herniation of spinal intervertebral disc with high accuracy and real-time. © 2020 IEEE.

关键词： Image segmentation

A Robust Remote Client Password Authentication Technique for Smart Cities E-Governance Applications

学校读者我要写书评

暂无评论

A Robust Remote Client Password Authentication Technique for...

2023 IEEE International Conference on Communications, Computing, Cybersecurity and Informatics, CCCI 2023

作者： Obaidat, Mohammad S. Meshram, Chandrashekhar Kartikey, Deepak Hsiao, Kuei-Fang Imoize, Agbotiname Lucky Barve, Sharad Kumar Abdullah Ii School of Information Technology The University of Jordan Amman11942 Jordan School of Computer and Communication Engineering University of Science and Technology Beijing Beijing100083 China School of Computing Srm University Department of Computational Intelligence TN Kattankulathur603203 India School of Engineering The Amity University UP Noida201301 India Jayawanti Haksar Government Post Graduation College College of Barkatullah University Department of Post Graduation Studies and Research in Mathematics M.P. Betul460001 India S.S.P. Govt. College Waraseoni Department of Mathematics M.P. Balaghat India University of Texas-Permian Basin Computer Science Department OdessaTX79762 United States University of Lagos Faculty of Engineering Department of Electrical and Electronics Engineering Lagos100213 Nigeria Institute of Digital Communication Ruhr University Department of Electrical Engineering and Information Technology Bochum44801 Germany Pandit S. N. Shukla University Shahdol Department of Mathematics Madhya Pradesh Shahdol484001 India

ISBN: (纸本)9798350302561

Smart cities are fast proliferating the global space in recent times. In order to achieve the envisioned next-generation cities, the smart city paradigm aspires to develop a completely integrated network-based infrastructure. With the development of innovative smart cities, ICT-based e-governance has made it possible for services to be delivered intelligently and interactively via online platforms. However, the communication between government agencies and other stakeholders in the smart city ecosystem is through open channels, which poses serious security concerns. In order to address this problem, this work proposes a robust remote client password authentication technique for e-governance applications in smart cities. The proposed authentication technique is lightweight, resistant to several attacks, and achieves low computational overhead. © 2023 IEEE.

关键词： Smart cards

Implantable silicon neural probes with nanophotonic phased arrays for single-lobe beam steering

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Chen, Fu-Der Sharma, Ankita Xue, Tianyuan Jung, Youngho Govdeli, Alperen Mak, Jason C.C. Chameh, Homeira Moradi Movahed, Mandana Brunk, Michael G.K. Luo, Xianshu Chua, Hongyao Lo, Patrick Guo-Qiang Valiante, Taufik A. Sacher, Wesley D. Poon, Joyce K.S. Max Planck Institute of Microstructure Physics Weinberg 2 Halle06120 Germany Department of Electrical and Computer Engineering University of Toronto 10 King’s College Road TorontoONM5S 3G4 Canada Max Planck-University of Toronto Centre for Neural Science and Technology Canada Krembil Brain Institute University Health Network TorontoON Canada Advanced Micro Foundry Pte. Ltd. 11 Science Park Road Singapore Science Park II 117685 Singapore Division of Neurosurgery Department of Surgery Toronto Western Hospital University of Toronto TorontoON Canada Institute of Biomedical Engineering University of Toronto TorontoON Canada

In brain activity mapping experiments using optogenetics, patterned illumination is crucial for deterministic and localized stimulation of neurons. However, due to optical scattering in brain tissue, light-emitting implantable devices are needed to bring precise patterned illumination to deep brain regions. A promising solution is silicon neural probes with integrated nanophotonic circuits that form tailored beam emission patterns without lenses. Here, we demonstrate neural probes with grating-based light emitters that generate a single steerable light beam across > 60% of the steering range with ≥ 4 dB of background suppression for optogenetic photostimulation. The light emitters, optimized for blue or amber light, combine end-fire optical phased arrays with slab gratings to suppress higher-order sidelobes. In vivo optogenetic photostimulation was demonstrated in Thy1-ChR2-YFP mice. These neural probes, capable of continuous single beam scanning, offer a promising avenue for optogenetic stimulation with cellular-level spatial precision in the deep brain. © 2024, CC BY-NC-ND.

关键词： Nanophotonics

QNN-VRCS: A Quantum Neural Network for Vehicle Road Cooperation Systems

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Innan, Nouhaila Behera, Bikash K. Al-Kuwari, Saif Farouk, Ahmed Quantum Physics and Magnetism Team LPMC Faculty of Sciences Ben M’sick Hassan II University of Casablanca Morocco United Arab Emirates NYUAD Research Institute NYUAD Abu Dhabi United Arab Emirates Pvt. Ltd. WB Mohanpur741246 India Qatar Center for Quantum Computing College of Science and Engineering Hamad Bin Khalifa University Qatar Foundation Doha Qatar Department of Computer Science Faculty of Computers and Artificial Intelligence South Valley University Hurghada Egypt

The escalating complexity of urban transportation systems, exacerbated by factors such as traffic congestion, diverse transportation modalities, and shifting commuter preferences, necessitates the development of more sophisticated analytical frameworks. Traditional computational approaches often struggle with the voluminous datasets generated by real-time sensor networks, and they generally lack the precision needed for accurate traffic prediction and efficient system optimization. This research integrates quantum computing techniques to enhance Vehicle Road Cooperation Systems (VRCS). By leveraging quantum algorithms, specifically UU† and variational UU†, in conjunction with quantum image encoding methods such as Flexible Representation of Quantum Images (FRQI) and Novel Enhanced Quantum Representation (NEQR), we propose an optimized Quantum Neural Network (QNN). This QNN features adjustments in its entangled layer structure and training duration to better handle the complexities of traffic data processing. Empirical evaluations on two traffic datasets show that our model achieves superior classification accuracies of 97.42% and 84.08% and demonstrates remarkable robustness in various noise conditions. This study underscores the potential of quantum-enhanced 6G solutions in streamlining complex transportation systems, highlighting the pivotal role of quantum technologies in advancing intelligent transportation solutions. © 2024, CC BY.

关键词： Traffic congestion

Reconfigurable Heterogeneous Parallel Island Models

学校读者我要写书评

暂无评论

Reconfigurable Heterogeneous Parallel Island Models

IEEE Symposium Series on Computational Intelligence (SSCI)

作者： Lucas A. Da Silveira Thaynara A. De Lima Mauricio Ayala-Rincón Department of Computer Science Universidade de Brasilia Brasilia D.F. Brazil Institute of Mathematics and Statistics Universidade Federal de Goiás - Campus II Goiânia Brazil Department of Mathematics Universidade de Brasilia Brasilia D.F. Brazil

ISBN: (纸本)9781665487696

Heterogeneous Parallel Island Models (HePIMs) run different bio-inspired algorithms (BAs) in their islands. From a variety of communication topologies and migration policies fine-tuned for homogeneous PIMs (HoPIMs), which run the same BA in all their islands, previous work introduced HePIMs that provided competitive quality solutions regarding the best-adapted BA in HoPIMs. This work goes a step forward, maintaining the population diversity provided by HePIMs, and increasing their flexibility, allowing BA reconfiguration on islands during execution: according to their performance, islands may substitute their BAs dynamically during the evolutionary process. Experiments with the introduced architectures were applied to the NP-hard problem of sorting permutations by reversals, using four different BAs, namely, simple Genetic Algorithm, Double-point crossover Genetic Algorithm, Differential Evolution, and self-adjusting Particle Swarm Optimization. The results showed that the novel reconfigurable heterogeneous models compute better quality solutions than the HePIMs closing the gap with the HoPIM running the best-adapted BA.

关键词： Barium NP-hard problem Computational modeling Sociology computer architecture Topology Statistics