检索结果-内蒙古大学图书馆

arXiv 2020年

作者： Hu, Hu Siniscalchi, Sabato Marco Wang, Yannan Bai, Xue Du, Jun Lee, Chin-Hui School of Electrical and Computer Engineering Georgia Institute of Technology United States Computer Engineering School University of Enna Italy Tencent Media Lab Tencent Corporation Shenzhen Guangdong China University of Science and Technology of China HeFei China

In this paper, we propose a sub-utterance unit selection framework to remove acoustic segments in audio recordings that carry little information for acoustic scene classification (ASC). Our approach is built upon a universal set of acoustic segment units covering the overall acoustic scene space. First, those units are modeled with acoustic segment models (ASMs) used to tokenize acoustic scene utterances into sequences of acoustic segment units. Next, paralleling the idea of stop words in information retrieval, stop ASMs are automatically detected. Finally, acoustic segments associated with the stop ASMs are blocked, because of their low indexing power in retrieval of most acoustic scenes. In contrast to building scene models with whole utterances, the ASM-removed sub-utterances, i.e., acoustic utterances without stop acoustic segments, are then used as inputs to the AlexNet-L back-end for final classification. On the DCASE 2018 dataset, scene classification accuracy increases from 68%, with whole utterances, to 72.1%, with segment selection. This represents a competitive accuracy without any data augmentation, and/or ensemble strategy. Moreover, our approach compares favourably to AlexNet-L with attention. Copyright © 2020, The Authors. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Face Photo-Sketch Synthesis Based on Conditional Adversarial Networks 16th

Face Photo-Sketch Synthesis Based on Conditional Adversarial...

引用

16th Web Information Systems and Applications Conference, WISA 2019

作者： Xie, Xiaoqi Xu, Huarong Weng, Lifen Electrical Engineering Xiamen University of Technology Xiamen China Computer Science and Technology Xiamen University of Technology Xiamen China Academy of Design Arts Xiamen University of Technology Xiamen China

ISBN: (纸本)9783030309510

Face Photo-Sketch synthesis is designed to generate an image contains rich personal information and details facial. It is widely used in law enforcement and life areas. Although some existing methods have achieved remarkable results, however, due to the gap between the synthetic and real image distribution, the synthetic image does not achieve the expected effect. To solve this problem, In this paper, we proposed conditional generation adversarial networks (CGAN) based on arts drawing steps constrain. We divide the process into two stages. In the first stage, we take the original face photos which concat noise z as input, generating stage 1 low resolution synthesize sketches. In the second stage take the stage 1 results and original face photo as inputs, yielding stage 2 high resolution synthesize sketches, which can express more natural textures and details. To add realism, we train our network using an adversarial loss. Experiments have shown that, Compared with previous methods, our results generate visually comfortable face sketches. And express more natural textures and details. © 2019, Springer Nature Switzerland AG.

关键词： Textures

来源：评论

学校读者我要写书评

暂无评论

A Two-Stage Approach to Device-Robust Acoustic Scene Classification

arXiv

引用

arXiv 2020年

作者： Hu, Hu Yang, Chao-Han Huck Xia, Xianjun Bai, Xue Tang, Xin Wang, Yajian Niu, Shutong Chai, Li Li, Juanjuan Zhu, Hongning Bao, Feng Zhao, Yuanjun Siniscalchi, Sabato Marco Wang, Yannan Du, Jun Lee, Chin-Hui School of Electrical and Computer Engineering Georgia Institute of Technology GA United States Tencent Media Lab Tencent Corporation China University of Science and Technology of China HeFei China Computer Engineering School University of Enna Kore Italy

To improve device robustness, a highly desirable key feature of a competitive data-driven acoustic scene classification (ASC) system, a novel two-stage system based on fully convolutional neural networks (CNNs) is proposed. Our two-stage system leverages on an ad-hoc score combination based on two CNN classifiers: (i) the first CNN classifies acoustic inputs into one of three broad classes, and (ii) the second CNN classifies the same inputs into one of ten finer-grained classes. Three different CNN architectures are explored to implement the two-stage classifiers, and a frequency sub-sampling scheme is investigated. Moreover, novel data augmentation schemes for ASC are also investigated. Evaluated on DCASE 2020 Task 1a, our results show that the proposed ASC system attains a state-of-the-art accuracy on the development set, where our best system, a two-stage fusion of CNN ensembles, delivers a 81.9% average accuracy among multi-device test data, and it obtains a significant improvement on unseen devices. Finally, neural saliency analysis with class activation mapping (CAM) gives new insights on the patterns learnt by our models. © 2020, CC-BY.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

DEVICE-ROBUST ACOUSTIC SCENE CLASSIFICATION BASED ON TWO-STAGE CATEGORIZATION AND DATA AUGMENTATION

arXiv

引用

arXiv 2020年

In this technical report, we present a joint effort of four groups, namely GT, USTC, Tencent, and UKE, to tackle Task 1 - Acoustic Scene Classification (ASC) in the DCASE 2020 Challenge. Task 1 comprises two different sub-tasks: (i) Task 1a focuses on ASC of audio signals recorded with multiple (real and simulated) devices into ten different fine-grained classes, and (ii) Task 1b concerns with classification of data into three higher-level classes using low-complexity solutions. For Task 1a, we propose a novel two-stage ASC system leveraging upon ad-hoc score combination of two convolutional neural networks (CNNs), classifying the acoustic input according to three classes, and then ten classes, respectively. Four different CNN-based architectures are explored to implement the two-stage classifiers, and several data augmentation techniques are also investigated. For Task 1b, we leverage upon a quantization method to reduce the complexity of two of our top-accuracy three-classes CNN-based architectures. On Task 1a development data set, an ASC accuracy of 76.9% is attained using our best single classifier and data augmentation. An accuracy of 81.9% is then attained by a final model fusion of our two-stage ASC classifiers. On Task 1b development data set, we achieve an accuracy of 96.7% with a model size smaller than 500KB 1. Copyright © 2020, The Authors. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Generic Tabu Search

引用

Journal of Software engineering and Applications 2022年第7期15卷 262-273页

作者： Chadi Kallab Samir Haddad Imad El-Zakhem Jinane Sayah Mohamad Chakroun Nisrine Turkey Jinan Charafeddine Hani Hamdan Wafaa Shakir Department of Computer Science and Mathematics Faculty of Arts and Sciences University of Balamand Koura Lebanon Department of Telecom and Networks Issam Fares Faculty of Technology University of Balamand Koura Lebanon Faculty of Computer Science and Electrical Engineering Universit&#228 t Rostock Rostock Germany Faculty of Engineering Notre Dame University Jounieh Lebanon Université Paris-Saclay P&#244 le scientifique et technologique de Vélizy Laboratoire d’Ingénierie des Systèmes de Versailles (LISV EA4048) Vélizy France Université Paris-Saclay CentraleSupélec CNRS Laboratoire des Signaux et Systèmes (L2S UMR CNRS 8506) Gif-sur-Yvette France Department of Computer Systems Faculty of Engineering Al-Furat Al-Awsat Technical University Babil Iraq

The Multiple Sequence Alignment problem is considered to be an NP-Hard problem, requiring initially a specific encoding schema and design, as for any other of its siblings, to implement and run any of the main categories of heuristic. This paper intends to discuss our proposed generic implementation of the Tabu Search algorithm, a heuristic procedure proposed by Fred Glover to solve discrete combinatorial optimization problems. In this research, we try to coordinate and synchronize different designs/implementations discussed in many literatures, with some of the references mentioned in this paper. The basic idea is to avoid that the search for best solutions stops when a local optimum is found, by maintaining a list of non-acceptable or forbidden (taboo) solutions/costs, called Tabu list or Short-Term Memory (STM). In our algorithm, we attempt to add some executions tracing functionalities in order to help later analysis for initial parameters tuning. On the other hand, we propose to include the concept of a list called Long-Term Memory (LTM), so that some of the best solutions found so far can be saved, for search diversification.

关键词： Generic Heuristics Bio-Informatics NP-Hard Tabu Search STM LTM

来源：评论

学校读者我要写书评

暂无评论

High-Gain Metasurface in Polyimide On-Chip Antenna Based on CRLH-TL for Sub-Terahertz Integrated Circuits (vol 10, 4298, 2020)

引用

SCIENTIFIC REPORTS 2020年第1期10卷 1-9页

作者： Alibakhshikenari, Mohammad Virdee, Bal S. See, Chan H. Abd-Alhameed, Raed A. Falcone, Francisco Limiti, Ernesto Electronic Engineering Department University of Rome 'Tor Vergata' Via del Politecnico 1 00133 Rome Italy London Metropolitan University Center for Communications Technology & Mathematics School of Computing & Digital Media N7 8DB London UK School of Engineering & the Built Environment Edinburgh Napier University 10 Colinton Road EH10 5DT Edinburgh UK School of Engineering University of Bolton Deane Road BL3 5AB Bolton UK School of Electrical Engineering & Computer Science University of Bradford BD7 1DP Bradford UK Electrical and Electronic Engineering Department Public University of Navarre 31006 Pamplona Spain

This paper presents a novel on-chip antenna using standard CMOS-technology based on metasurface implemented on two-layers polyimide substrates with a thickness of 500 μm. The aluminium ground-plane with thickness of 3 μm is sandwiched between the two-layers. Concentric dielectric-rings are etched in the ground-plane under the radiation patches implemented on the top-layer. The radiation patches comprise concentric metal-rings that are arranged in a 3 × 3 matrix. The antennas are excited by coupling electromagnetic energy through the gaps of the concentric dielectric-rings in the ground-plane using a microstrip feedline created on the bottom polyimide-layer. The open-ended feedline is split in three-branches that are aligned under the radiation elements to couple the maximum energy. In this structure, the concentric metal-rings essentially act as series left-handed capacitances CL that extend the effective aperture area of the antenna without affecting its dimensions, and the concentric dielectric rings etched in the ground-plane act as shunt left-handed inductors LL, which suppress the surface-waves and reduce the substrates losses that leads to improved bandwidth and radiation properties. The overall structure behaves like a metasurface that is shown to exhibit a very large bandwidth of 0.350–0.385 THz with an average radiation gain and efficiency of 8.15dBi and 65.71%, respectively. It has dimensions of 6 × 6 × 1 mm3 that makes it suitable for on-chip implementation. [ABSTRACT FROM AUTHOR]

关键词：

来源：评论

学校读者我要写书评

暂无评论

Elastowave: Localized Tactile Feedback in a Soft Haptic Interface via Focused Elastic Waves

Elastowave: Localized Tactile Feedback in a Soft Haptic Inte...

引用

International Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Systems (HAPTICS)

作者： Gregory Reardon Nikolas Kastor Yitian Shao Yon Visell Media Arts and Technology Program University of California Santa Barbara California California NanoSystems Institute University of California Santa Barbara Santa Barbara California Department of Electrical and Computer Engineering University of Santa Barbara Santa Barbara California

ISBN: (数字)9781728102344

ISBN: (纸本)9781728102351

It is challenging to engineer programmable tactile displays to match human haptic abilities. Such displays are often composed of elements whose stiffness contrasts greatly with the softness of many natural materials. Emerging soft material technologies hold promise for their ability to conform to many objects, including the human body. However, rendering localized feedback from soft haptic devices remains challenging. Here, we present the Elastowave, a soft tactile interface that provides localized tactile feedback via a soft, compliant surface. We achieve this by focusing elastic wave fields generated by a compact array of remotely-positioned actuators. Our method is based on new variations of time-reversal focusing techniques for elastodynamic waves. Our system can provide dynamic, single- or multi-point localized tactile feedback with centimeter-scale resolution across a deformable interface with an area of 175 cm 2 . The sizeable displacements of the focused tactile signals enable them to be easily felt, as our experiments show. This work could enable the design of a multitude of new soft tactile interfaces in areas such as creative computing, product design, and augmented reality.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Non-invasive suppression of essential tremor via phase-locked disruption of its temporal coherence

引用

Brain Stimulation 2023年第1期16卷 276-277页

作者： Li, Junheng Schreglmann, Sebastian Wang, David Peach, Robert Laurate, Anna Rhodes, Edward Panella, Emanuele Cassara, Antonino Boyden, Edward Barahona, Mauricio Rothwell, John Bhatia, Kailash Grossman, Nir Department of Brain Sciences Imperial College London London UK UK Dementia Research Institute UK Institute of Neurology Department of Clinical and Movement Neuroscience Queen Square University College London (UCL) UK Computer Science and Artificial Intelligence Laboratory Massachussetts Institute of Technology (MIT) USA Department of Brain Sciences Imperial College London UK Department of Mathematics and EPSRC Centre for Mathematics of Precision Healthcare Imperial College London UK Department of Physics Imperial College London London SW7 2AZ UK IT’IS Foundation for Research on Information Technologies in Society Switzerland Department of Media Arts and Sciences MIT USA McGovern Institute for Brain Research MIT USA Howard Hughes Medical Institute Cambridge USA Department of Biological Engineering MIT USA Department of Brain and Cognitive Sciences MIT USA Centre for Neurobiological Engineering MIT USA Koch Institute for Integrative Cancer Research MIT USA Centre for Bio-Inspired Technology Department of Electrical and Electronic Engineering Imperial College London UK Centre for Neurotechnology Imperial College London UK

来源：评论

学校读者我要写书评

暂无评论

Gamification, sensory feedback, adaptive function on virtual reality rehabilitation: a brief review

Gamification, sensory feedback, adaptive function on virtual...

引用

Digital Transformation and Intelligence (ICDI), International Conference on

作者： Fatin Shamimi Mohd Zuki Frédéric Merienne Suziah Sulaiman Aylen Ricca Dayang Rohaya Awang Rambli Mohamad Naufal Mohamad Saad Arts et Metiers Institute of Technology LISPEN HESAM Université Chalon-Sur-Saône France Computer and Information Sciences Department Universiti Teknologi PETRONAS Seri Iskandar Malaysia Département Automatique Productique et Informatique IMT Atlantique LS2N-CNRS Nantes France Electrical and Electronic Engineering Department Universiti Teknologi PETRONAS Seri Iskandar Malaysia

ISBN: (纸本)9798350397017

Feedback is often linked to rewards element in gamification to motivate users. However, there are more to feedbacks than rewards elements. Feedback can provide user with useful information and at the same time feedback from user are useful for an adaptive system. There has been lack of empirical basis for feedback design to maintain motivation in virtual reality environment especially for the less cognitively abled population such as stroke patients. This paper presents a review study on gamification, sensory feedback, and adaptive function with regards to the use of virtual reality in stroke rehabilitation. A further analysis was conducted on those work that involves gamification and virtual reality design principles. One major finding is the dominant use of visual feedback in the design as compared to auditory and haptics despite their potentials to encourage motivation and engagement. The literature findings will be used to inform future empirical research on virtual reality design for stroke rehabilitation. The idea is to investigate the effect of sensory feedback and the added value on motivation and engagement. Some plans on how to conduct such a study will be illustrated.

关键词： Visualization Adaptive systems Digital transformation Sociology Virtual reality Stroke (medical condition) Haptic interfaces

来源：评论

学校读者我要写书评

暂无评论

Retraction Note to: Threshold based DDoS mitigation with fog layer in cloud environment

引用

Journal of Ambient Intelligence and Humanized Computing 2022年第1期14卷 669-669页

作者： Dhiyanesh, B. Sakthivel, S. Radha, R. Kumar, S. Senthil Department of Computer Science and Engineering Hindusthan College of Engineering and Technology Coimbatore India Department of Computer Science and Engineering Sona College of Technology Salem India Department of Electrical and Electronics Engineering Karpagam Institute of Technology Coimbatore India Department of Commerce With Computer Applications Dr. SNS Rajalakshmi College of Arts and Science Coimbatore India

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：