检索结果-内蒙古大学图书馆

Monaural speech enhancement using U-net fused with multi-head self-attention

Chinese Journal of Acoustics 2023年第1期42卷 98-118页

作者： FAN Junyi YANG Jibin ZHANG Xiongwei ZHENG Changyan Graduate School Army Engineering UniversityNanjing 210007 College of Command and Control Engineering Army Engineering UniversityNanjing 210007 Department of Test Control High-tech InstituteQingzhou 262500

Under low signal-to-noise ratio(SNR)and burst noise conditions,the speech enhancement effect of existing deep learning network models is not *** contrast,humans can exploit the long-term correlation of speech to form an integrated perception of different speech ***,describing the long-term dependencies of speech can help improve the enhancement performance under low SNR and burst noise *** by this feature,a time domain end-to-end monaural speech enhancement model TU-net that fuses the multi-head self-attention mechanism and U-net deep network is *** TU-net model adopts the codec layer structure of U-net to achieve multi-scale feature *** introduces the dual-path Transformer module using the multi-head self-attention mechanism to calculate the speech mask and better model long-term *** TU-net model is trained with a weighted sum loss function in the time,time-frequency,and perceptual *** experiments are carried out and the results show that with maintaining relatively fewer network model parameters,TU-net outperforms other similar monaural enhancement network models in several evaluation metrics such as perceptual evaluation of speech quality(PESQ),short-time objective intelligibility(STOI)and SNR gain under low SNR and burst noise conditions.

关键词： network speech noise

来源：评论

学校读者我要写书评

暂无评论

An Energy Stable Bound-Preserving Finite Volume Scheme for the Allen-Cahn Equation Based on Operator Splitting Method

SSRN

引用

SSRN 2024年

作者： Peng, Gang Li, Yuan Command and Control Engineering College Army Engineering University of PLA Nanjing China Department of Basic Education Army Engineering University of PLA Nanjing China

In this paper, an energy stable bound-preserving finite volume scheme is constructed for the Allen-Cahn equation. The first-order operator splitting method is used to split the original equation into a nonlinear equation and a heat equation in each time interval. The nonlinear equation is solved by the backward Euler scheme, and the heat equation is discretized by the extremum-preserving scheme. The harmonic averaging points on cell facets are employed to define auxiliary unknowns, which enable our discrete scheme to be applicable to unstructured meshes. The energy stable and bound-preserving analysis of the finite volume scheme is also presented. Numerical experiments illustrate that this linear numerical scheme is practical and accurate in solving the Allen-Cahn equation. © 2024, The Authors. All rights reserved.

关键词： Nonlinear equations

来源：评论

学校读者我要写书评

暂无评论

Paying more attention to local contrast: improving infrared small target detection performance via prior knowledge

arXiv

引用

arXiv 2024年

作者： Wang, Peichao Wang, Jiabao Chen, Yao Zhang, Rui Li, Yang Miao, Zhuang Command and Control Engineering College Army Engineering University of PLA Jiangsu Nanjing210017 China

The data-driven method for infrared small target detection (IRSTD) has achieved promising results. However, due to the small scale of infrared small target datasets and the limited number of pixels occupied by the targets themselves, it is a challenging task for deep learning methods to directly learn from these samples. Utilizing human expert knowledge to assist deep learning methods in better learning is worthy of exploration. To effectively guide the model to focus on targets’ spatial features, this paper proposes the Local Contrast Attention Enhanced infrared small target detection Network (LCAE-Net), combining prior knowledge with data-driven deep learning methods. LCAE-Net is a U-shaped neural network model which consists of two developed modules: a Local Contrast Enhancement (LCE) module and a Channel Attention Enhancement (CAE) module. The LCE module takes advantages of prior knowledge, leveraging handcrafted convolution operator to acquire Local Contrast Attention (LCA), which could realize background suppression while enhance the potential target region, thus guiding the neural network to pay more attention to potential infrared small targets’ location information. To effectively utilize the response information throughout downsampling progresses, the CAE module is proposed to achieve the information fusion among feature maps’ different channels. Experimental results indicate that our LCAE-Net outperforms existing state-of-the-art methods on the three public datasets NUDT-SIRST, NUAA-SIRST, and IRSTD-1K, and its detection speed could reach up to 70 fps. Meanwhile, our model has a parameter count and Floating-Point Operations (FLOPs) of 1.945M and 4.862G respectively, which is suitable for deployment on edge devices. Copyright © 2024, The Authors. All rights reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Abnormal data detection method based on ant colony clustering 22

Abnormal data detection method based on ant colony clusterin...

引用

Proceedings of the 2022 6th International Conference on Electronic Information Technology and Computer engineering

作者： Lianyong Chen Jinyu Song College of Command and Control Engineering Army Engineering University of PLA China

ISBN: (纸本)9781450397148

With the rapid growth of data, low quality data emerges, and data quality problems are increasingly serious. Abnormal data detection becomes an important aspect of data quality validity measurement. We propose an improved ant colony clustering algorithm based on LF algorithm [8] by optimizing ant position, attribute weight and moving step speed and then apply it to the abnormal data detection. The improved algorithm is programmed by Python, and the experiment is carried out on UCI data set. The results show that the improved algorithm can obtain better clustering effect, and make the accuracy of abnormal data detection results higher.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Evolution Mechanism of Unmanned Cluster Cooperation Oriented Toward Strategy Selection Diversity

引用

Journal of Systems engineering and Electronics 2025年第2期36卷 462-482页

作者： Xie Zhenhai Yu Minggang He Ming Chen Guoyou Zhai Zheng Wang Ziyu Liu Lu Institute of Command and Control Engineering Army Engineering University of PLA Nanjing China Unit 31630 of the Chinese People's Liberation Army Huizhou China

When performing tasks, unmanned clusters often face a variety of strategy choices. One of the key issues in unmanned cluster tasks is the method through which to design autonomous collaboration and cooperative evolution mechanisms that allow for unmanned clusters to maximize their overall task effectiveness under the condition of strategic diversity. This paper analyzes these task requirements from three perspectives: the diversity of the decision space, information network construction, and the autonomous collaboration mechanism. Then, this paper proposes a method for solving the problem of strategy selection diversity under two network structures. Next, this paper presents a Moran-rule-based evolution dynamics model for unmanned cluster strategies and a vision-driven-mechanism-based evolution dynamics model for unmanned cluster strategy in the context of strategy selection diversity according to various unmanned cluster application scenarios. Finally, this paper provides a simulation analysis of the effects of relevant parameters such as the payoff factor and cluster size on cooperative evolution in autonomous cluster collaboration for the two types of models. On this basis, this paper presents advice for effectively addressing diverse choices in unmanned cluster tasks, thereby providing decision support for practical applications of unmanned cluster tasks.

关键词： Games Collaboration Modeling Stochastic processes Information processing Heuristic algorithms Complex networks Investment Hands Decision making

来源：评论

学校读者我要写书评

暂无评论

Effect of Neutrinos on Angular Momentum of Dark Matter Halo

引用

Research in Astronomy and Astrophysics 2023年第8期23卷 355-363页

作者： Yu Chen Chang-Zhi Lu Yu Lu Tingting Zhang Tong-Jie Zhang Institute for Frontiers in Astronomy and Astrophysics Beijing Normal UniversityBeijing 102206China Department of Astronomy Beijing Normal UniversityBeijing 100875China Beijing Planetarium Beijing Academy of Science and TechnologyBeijing 100044China College of Command and Control Engineering PLA Army Engineering UniversityNanjing 210017China Institute for Astronomical Science Dezhou UniversityDezhou 253023China

Massive neutrinos are expected to affect the large-scale structure formation,including the major component of solid substances,dark matter *** halos are influenced by neutrinos is vital and interesting,and angular momentum(AM)as a significant feature provides a statistical perspective for this *** halos from TianNu N-body cosmological simulation with the co-evolving neutrino particles,we obtain some concrete ***,by comparing the same halos with and without neutrinos,in contrast to the neutrino-free case,over 89.71%of halos have smaller halo moduli,over 71.06%have smaller particle-mass-reduced(PMR)AM moduli,and over 95.44%change their orientations of less than 0°.***,the relative variation of PMR modulus is more visible for low-mass ***,to explore the PMR moduli of halos in dense or sparse areas,we divide the whole box into big cubes,and search for halos within a small spherical cell in a single *** the two-level divisions,we discover that in denser cubes,the variation of PMR moduli with massive neutrinos decreases more *** distinction suggests that neutrinos exert heavier influence on halos'moduli in compact *** massive neutrinos,most halos(86.60%)have lower masses than without neutrinos.

关键词： (cosmology:)dark matter-neutrinos-(cosmology:)large-scale structure of universe-methods data analysis

来源：评论

学校读者我要写书评

暂无评论

FewSAR: A Few-shot SAR Image Classification Benchmark

arXiv

引用

arXiv 2023年

作者： Zhang, Rui Wang, Ziqi Li, Yang Wang, Jiabao Wang, Zhiteng Command & Control Engineering College Army Engineering University of PLA Nanjing China

Few-shot learning (FSL) is one of the significant and hard problems in the field of image classification. However, in contrast to the rapid development of the visible light dataset, the progress in SAR target image classification is much slower. The lack of unified benchmark is a key reason for this phenomenon, which may be severely overlooked by the current literature. The researchers of SAR target image classification always report their new results on their own datasets and experimental setup. It leads to inefficiency in result comparison and impedes the further progress of this area. Motivated by this observation, we propose a novel few-shot SAR image classification benchmark (FewSAR) to address this issue. FewSAR consists of an open-source Python code library of 15 classic methods in three categories for few-shot SAR image classification. It provides an accessible and customizable testbed for different few-shot SAR image classification task. To further understanding the performance of different few-shot methods, we establish evaluation protocols and conduct extensive experiments within the benchmark. By analyzing the quantitative results and runtime under the same setting, we observe that the accuracy of metric learning methods can achieve the best results. Meta-learning methods and fine-tuning methods perform poorly on few-shot SAR images, which is primarily due to the bias of existing datasets. We believe that FewSAR will open up a new avenue for future research and development, on real-world challenges at the intersection of SAR image classification and few-shot deep learning. We will provide our code for the proposed FewSAR at https://***/solarlee/FewSAR. © 2023, CC BY-NC-ND.

关键词： Synthetic aperture radar

来源：评论

学校读者我要写书评

暂无评论

Deep Hashing Retrieval of Dynamic Corpora

Deep Hashing Retrieval of Dynamic Corpora

引用

6th International Conference on Intelligent Computing and Signal Processing (ICSP)

作者： Ao Zou Wenning Hao Tianhao Shao Ke Zhang Qiuhui Shen College of Command & Control Engineering Army Engineering University of PLA Nanjing China

Knowledge-intensive language applications, like open-domain question answering, call for NLP systems to retrieve related evidence in a document collection. A efficient and scalable retrieval system is an ideal component in these applications. Up-to-data research developed sequence-to-sequence framework on this task, eschewing any format of term-frequency or query-document similarity calculation. This line of work greatly improved the efficient of retrieval but with the sacrifice of scalablity. This work proposes a new framework based on the seq2seq architecture, which can solve scalability issues that exist in such methods. Experimental results on the public dataset Natural Question demonstrate the effectiveness of this approach.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Android-Based Binary Code Function Fast Positioning Technology

Android-Based Binary Code Function Fast Positioning Technolo...

引用

Information Technology,Big Data and Artificial Intelligence (ICIBA), International Conference on IEEE International Conference on

作者： Xingyi Guan Min Zhao Command and Control Engineering College Army Engineering University of PLA Nanjing China

To solve the problem that the existing Android reverse analysis method is inefficient and the applicable scenario is single. A fast positioning technology for binary code function based on Android is proposed in this paper, which includes preprocessing, behavior analysis combining dynamic and static, and binary code function comparison, which can accurately and quickly locate the location of binary code function. Based on this research, to further broaden the application scenarios of the fast positioning technology proposed in this paper and realize the automatic analysis of large-scale applications, an automated analysis tool AndPosit is designed. The proposed technology has been experimentally verified that it can be widely used in various Android applications, can accurately locate the target function code, and the efficiency is more than dozens of times higher than the original manual analysis method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Exploration of a Blended Teaching Mode for the Cultivation of High-Order Computing Ability 14

Exploration of a Blended Teaching Mode for the Cultivation o...

引用

14th International Conference on Education Technology and Computers, ICETC 2022

作者： Wang, Meijuan Wang, Yibin Li, Hui College of Command and Control Engineering The Army Engineering University of PLA Nanjing210001 China College of Computer and Information Anqing Normal University Anqing246133 China

ISBN: (纸本)9781450397766

The rapid development of information technology has a great influence on the future society. Future talents should have high-order thinking skills, such as adaptability, competence, and innovation, suitable for specific positions. Military universities undertake the important task of cultivating military talent. This work highlights the cultivation of students' computing thinking ability in the course "The Fundamentals of Computer."By carefully planning teaching time and content and using modern information technology and rich network teaching resources, the reform of online-offline blended teaching is practiced. The new teaching mode features high-order thinking skills training for students and realizes the transformation of the curriculum from knowledge output to high-order thinking ability orientation. © 2022 ACM.

关键词： Students

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：