检索结果-内蒙古大学图书馆

Adversarial examples with transferred camouflage style for object detection

Journal of Physics: Conference Series 2021年第1期1738卷

作者： Xiaotong Deng Zheng Fang Yunfei Zheng Yang Wang Junming Huang Xiaobing Wang Tieyong Cao Institute of Command and Control Engineering Army Engineering University Nanjing Jiangsu 210000 China PLA Army Academy of Artillery and Air Defense Nanjing Jiangsu 210000 China

Most of existing adversarial examples attacking methods for object detection models aim at generating subtle perturbation which is invisible to human vision. However, some perturbations with stronger intensity are equally effective and can even make the objects in adversarial examples more invisible. In this paper, an attack model based on the Generative Adversarial Network and style transfer method is designed to craft and add perturbation with camouflage style to the object area in the image. Experiments on PASCAL VOC 2012 dataset demonstrated that Adversarial examples generated from our model can attack both proposal-based detectors and regression-based object detection models effectively, and reduce the visual saliency of the objects.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Research on Evaluation Method of Wargame Strategy Based on Fuzzy Petri Net 2

Research on Evaluation Method of Wargame Strategy Based on F...

引用

2nd IEEE International Conference on Information Systems and Computer Aided Education, ICISCAE 2019

作者： Shang, Tongfei Dong, Haoyang Han, Kun Li, Gang National University of Defense Technology College of Information and Communication Xi'an China School of Electronic and Information Engineering Xi'an Jiaotong University Xi'an China Army Engineering University of PLA College of Command and Control Engineering Nanjing China Xi'An Satellite Control Center Xi'an China

ISBN: (纸本)9781728130668

Wargame can effectively improve the fidelity and efficiency of combat deduction, and has the ability to solve complex military problems. It is theoretical and military significance to carry out the research on the development and application of computer wargame system. In this paper, the method of knowledge representation and reasoning of fuzzy Petri nets is applied to the strategy of wargame deduction. The characteristics of the graphs are used to represent different variables with different symbols, which can form a concise PN map. This PN map can be used for the force representation and reasoning strategy in the wargame system. The results show that the Petri net-based knowledge representation and reasoning algorithm is very effective for the wargame strategy. © 2019 IEEE.

关键词： Petri nets

来源：评论

学校读者我要写书评

暂无评论

CapsHash: Deep Supervised Hashing with Capsule Network

CapsHash: Deep Supervised Hashing with Capsule Network

引用

International Conference on Wireless Communications and Signal Processing (WCSP)

作者： Yang Li Rui Zhang Zhuang Miao Jiabao Wang Command & Control Engineering College Army Engineering University of PLA Nanjing China

To better deal with large-scale image retrieval problem, deep hashing models based on convolutional neural network (CNN) have been widely used as effective methods, which can map similar images to compact binary hash codes with smaller hamming distance. Despite their positive results, CNN-based methods have few limitations, which are unable to understand the spatial relationship between features. To overcome this challenge, in this paper, a novel deep supervised hashing method was proposed based on capsule networks. Our method, referred to as CapsHash, can learn discriminative hash codes and capsule vectors at the same time. Moreover, we introduce a novel compound loss function that has two parts: classification hashing loss and margin loss. This compound loss function can greatly improve the discriminative ability of binary codes and further improve the image retrieval performance. Extensive experiments under different scenarios demonstrate that our CapsHash method can preserve the instance-level similarity and outperform previous state-of-the-art hashing approaches. To the best of our knowledge, CapsHash is the first method about the application of capsule networks in the deep supervised hashing domain.

关键词： Image retrieval Training Compounds Semantics Computer architecture Neurons Binary codes

来源：评论

学校读者我要写书评

暂无评论

Research on Teaching Presence Evaluation Indexes of Online Course

Research on Teaching Presence Evaluation Indexes of Online C...

引用

IEEE International Symposium on Information (IT) in Medicine and Education, ITME

作者： Wenyu Zhang Weiwei Chen Honghua Zhao Rui Wang Department of command and control engineering Army Engineering University of PLA Nanjing China

Online courses are short of face-to-face and direct communication compared with classroom teaching, and the in-depth communication between teachers and students can only occur in the environment of high level of teaching presence. Therefore, how to establish and evaluate the presence in online teaching is the research objective. For the evaluation of teaching presence of online teaching, two-level evaluation index system is set up in a hierarchical way in our research. The first-level evaluation indexes include design organization; facilitate discussion and direct guidance. The primary evaluation index is refined in the second level evaluation indexes. The SPOC course “computer composition principle” is evaluated according to the two-level evaluation indexes. By analyzing the result, it can be concluded that the teaching presence evaluation indexes can provide effective references for teachers to improve teaching methods and teaching design, and form closed-loop feedback.

关键词： Education Indexes Organizations Videos Design methodology Task analysis Media

来源：评论

学校读者我要写书评

暂无评论

Collaborative Filtering Recommendation Algorithm Based on Random Forest Filling

Collaborative Filtering Recommendation Algorithm Based on Ra...

引用

International Conference on Information Systems and Computer Aided Education (ICISCAE)

作者： Zhongwei Wang Hangping Qiu Yi Sun Qiaoyu Deng School of Command and Control Engineering Army Engineering University of PLA Nanjing China

ISBN: (数字)9781728130668

ISBN: (纸本)9781728130675

Aiming at the sparsity problem of data set in the field of information recommendation, a collaborative filtering recommendation algorithm based on random forest filling is proposed. First, ID3 algorithm is adopted to construct the decision trees and the random forest is composed of the decision trees. Then, according to the steps of the proposed filling algorithm, the missing values of the data set are predicted using random forest and then be filled. Besides, the mean and the mode filling methods are used for a contrast. Finally, the collaborative filtering recommendation algorithm is tested for the datasets processed by different methods. Experimental results show that compared with the general recommendation algorithm and the recommendation algorithm based on the mean and the mode filling, the proposed recommendation algorithm based on random forest filling significantly improves the F value of the recommendation result and further guarantees the quality of the recommendation result.

关键词： Forestry Decision trees Prediction algorithms Collaboration Filtering Filling Training

来源：评论

学校读者我要写书评

暂无评论

The Detection Algorithms for Similar Duplicate Data

The Detection Algorithms for Similar Duplicate Data

引用

International Conference on Systems and Informatics (ICSAI)

作者： Jin-yu Song Quan Yu Ruo-yu Bao College of Command and Control Engineering Army Engineering University of PLA Nanjing China

ISBN: (数字)9781728152561

ISBN: (纸本)9781728152578

This paper studied and analyzed three algorithms which can be used to detect similar duplicate data records. Among them, the two commonly used duplicate data detection algorithms are basic sorted-neighborhood method (SNM) and priority queue algorithm. Both of them are based on sorting-merger thought. The third one is Density-Based Spatial Clustering of Applications with Noise (DBSCAN). The paper also discussed some key problems in the application and implementation of algorithms, including how to choose the attributes of data records and calculate the similarity between these attributes of two records, the choice of the sort algorithm, the setting of sliding window size and queue length. The paper realized three algorithms by programming based on Matlab platform, as well as some technologies related to algorithms. The paper verified related conclusions and evaluated three algorithms by testing a lot of loaded data sets. The paper finally provided the instruction to apply these algorithms.

关键词： Clustering algorithms Sorting Detection algorithms Eigenvalues and eigenfunctions Shape Measurement Data integrity

来源：评论

学校读者我要写书评

暂无评论

The Data Integrity Error Repair Method for Filling Missing Values with External Data

The Data Integrity Error Repair Method for Filling Missing V...

引用

International Congress on Image and Signal Processing, BioMedical engineering and Informatics (CISP-BMEI)

作者： Jinyu Song Quan Yu Yiping Guo College of Command and Control Engineering Army Engineering University of PLA Nanjing China

In this paper, the availability of data in a big data environment is used as an application requirement. The general method of populating non-inferential missing values based on function dependencies on external data is improved. A method of using external data to fill in missing values is proposed. Types of network data to query, each query has a trust value, multiple queries may have the same result, a trust value combination mechanism is used to calculate the trust value of the query result according to the trust value of the query, and execute as few as possible The query will use the most likely correct (highest trust value) query result as the final fill value. The experimental results show that the query construction method proposed in this paper, as well as the optimization and strategy selection of the trust value calculation method, can effectively improve the efficiency of query based on a large amount of external data.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Tiny Fusion: Tiny Deep Convolutional Neural Network for Real-time Image Fusion

Tiny Fusion: Tiny Deep Convolutional Neural Network for Real...

引用

IEEE International Conference on Information Communication and Signal Processing (ICICSP)

作者： Jixiao Wang Yang Li Zhuang Miao Jiabao Wang Xun Zhao Command & Control Engineering College Army Engineering University of PLA Nanjing China

Since convolutional neural networks (CNNs) were used in image fusion field, they have showed state-of-the-art quality beyond traditional methods. However, the existing CNN fusion models have a high computational cost and require high memory capacity, which is impractical for embedded applications or mobile platforms. Inspired by the efficiency of the lightweight network architecture of SqueezeNet, MobileNet, and ShuffleNet, we propose a tiny fusion method for image fusion, which significantly decreasing the number of operations and memory needed while retaining the same fusion quality. Extensive experimental results indicate that tiny deep neural network architectures can be designed for real-time image fusion that are well suited for embedded scenarios. To the best of our knowledge, our method is the first lightweight network architecture for image fusion.

关键词： Image fusion Feature extraction Network architecture Image reconstruction Computational modeling Computer architecture Deep learning

来源：评论

学校读者我要写书评

暂无评论

Likelihood-free cosmological constraints with artificial neural networks: An application on hubble parameters and SNe Ia

arXiv

引用

arXiv 2020年

作者： Wang, Yu-Chen Xie, Yuan-Bo Zhang, Tong-Jie Huang, Hui-Chao Zhang, Tingting Liu, Kun Department of Physics Beijing Normal University Beijing100875 China Department of Astronomy Beijing Normal University Beijing100875 China Institute for Astronomical Science Dezhou University Dezhou253023 China College of Command and Control Engineering PLA Army Engineering University Nanjing China

The errors of cosmological data generated from complex processes, such as the observational Hubble parameter data (OHD) and the Type Ia supernova (SN Ia) data, cannot be accurately modeled by simple analytical probability distributions, e.g. Gaussian distribution. To constrain cosmological parameters from these data, likelihood-free inference is usually used to bypass the direct calculation of the likelihood. In this paper, we propose a new procedure to perform likelihood-free cosmological inference using two artificial neural networks (ANN), the Masked Autoregressive Flow (MAF) and the denoising autoencoder (DAE). Our procedure is the first to use DAE to extract features from data, in order to simplify the structure of MAF needed to estimate the posterior. Tested on simulated Hubble parameter data with a simple Gaussian likelihood, the procedure shows the capability of extracting features from data and estimating posterior distributions without the need of tractable likelihood. We demonstrate that it can accurately approximate the real posterior, achieve performance comparable to the traditional MCMC method, and the MAF gets better training results for small number of simulation when the DAE is added. We also discuss the application of the proposed procedure to OHD and Pantheon SN Ia data, and use them to constrain cosmological parameters from the non-flat ΛCDM model. For SNe Ia, we use fitted light curve parameters to find constraints on H0,Ωm,ΩΛ similar to relevant work, using less empirical distributions. In addition, this work is also the first to use Gaussian process in the procedure of OHD simulation. Copyright © 2020, The Authors. All rights reserved.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

Feature Decomposition and Attention-guided Boundary Refinement for Saliency Detection

Feature Decomposition and Attention-guided Boundary Refineme...

引用

IEEE Advanced Information Management,Communicates,Electronic and Automation control Conference (IMCEC)

作者： Shengkai Xiang Tieyong Cao Shizhan Hong Zhen Liu Artificial Intelligence Lab Army Engineering University of PLA Nanjing China College of Command and Control Army Engineering University of PLA Nanjing China

In this paper, we propose a novel saliency detection model. We manage to balance the conflict between quality and efficiency by introducing Octave Convolution (OctConv). The proposed detection model is based on a typical encoder- decoder framework, and all encoded side-out features are de- composed into low and high frequency parts, respectively. The low frequency parts locate coarse salient regions, while the high frequency parts help capture edge details. And a two-stream attention module is proposed to emphasize important global and local context. A top-down attention transfers a global saliency prior to shallow layers and a bottom-up attention helps deep layers focus more on the edges. Compared to the state-of-the- art saliency methods, proposed model is of much lighter-weight (68.6 MB), runs faster (63.7 fps in real-time) and achieves competitive performances on six public benchmarks in terms of both quantitative and qualitative evaluation.

关键词：

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：