检索结果-内蒙古大学图书馆

A Text detection Algorithm for Image of Student Exercises Based on CTPN and Enhanced YOLOv3

IEEE ACCESS 2020年 8卷 176924-176934页

作者： Cao, Langcai Li, Hongwei Xie, Rongbiao Zhu, Jinrong Xiamen Univ Dept Automat Xiamen 361005 Peoples R China Xiamen Key Lab Big Data Intelligent Anal & Decis Xiamen 361005 Peoples R China

Intelligent learning system (ILS) has become a popular learning tool for students. It can collect students' wrong questions in exercises and dig out their unskilled knowledge points so that it can recommend personalized exercises for students. Detecting text accurately from images of students' exercises is significant and essential in an ILS. However, a big challenge of text detection is that traditional text detection algorithms can not detect complete text lines in an exercise scene, and their detection box always splits between Chinese and mathematical symbols. In this article, we propose a deep-learning-based approach for text detection, which improves You Only Look Once version 3 (YOLOv3) by changing the regression object from a single character to a fixed-width text and applies a stitching strategy to construct text lines based on the relation matrix, which improves the accuracy by 9.8%. Experimental results on both RCTW Chinese text detection dataset and real exercise scenario show that our model can improve detection effectiveness. In addition, we compare our method with two state-of-the-art approaches in applications of exercise text detection, and discuss its capability and limitations. We have also provided a platform which has implemented the proposal for detecting text lines in students' daily homework or examination papers, which enhances user experience well.

关键词： Text recognition Databases Splicing Detectors Feature extraction Object detection detection algorithms Text detection exercise image YOLOv3 CTPN OCR platform stitching algorithm

来源：评论

学校读者我要写书评

暂无评论

A uniformly most powerful detector of gas plumes against a cluttered background

引用

REMOTE SENSING OF ENVIRONMENT 2021年 260卷 112443-112443页

作者： Schaum, Alan US Naval Res Lab Washington DC 20375 USA

The agriculture, waste management, and energy sectors make major contributions of methane to the atmosphere. Cost-effective means of locating both anthropogenic and natural sources of methane and other gases are needed in order to improve understanding of greenhouse gas budgets and facilitate mitigation. One approach uses imaging spectrometers operating on remote sensing platforms, such as aircraft or satellites, to locate specific types of plumes. This paper develops spectral algorithms for the detection of any absorptive gas, for example, methane in the shortwave infrared region. The presence of a target gas is inferred from its known absorptivity spectrum along with local background statistics gathered by the sensor. The problem of distinguishing the target signal from clutter is usually solved by modeling the spectral background with some member of the class of elliptical distributions, most commonly a multivariate Gaussian. That class of models is replaced here by the lognormal distribution. A logarithmic transformation then recasts the associated hypothesis test into a standard form, allowing derivation of a uniformly most powerful solution. It is theoretically optimal simultaneously for all gas concentrations. This leads further to new algorithms for the detection and quantification of gas mixtures.

关键词： Imaging spectrometry Remote sensing Decision theory detection algorithms Gas plumes Uniformly most powerful detector Estimation Methane

来源：评论

学校读者我要写书评

暂无评论

On the Robustness of Random Walk algorithms for the detection of Unstructured P2P Botnets 11

On the Robustness of Random Walk Algorithms for the Detectio...

引用

11th International Conference on IT Security Incident Management and IT Forensics (IMF)

作者： Muhs, Dominik Haas, Steffen Strufe, Thorsten Fischer, Mathias Tech Univ Dresden Dresden Germany Univ Hamburg Hamburg Germany

ISBN: (纸本)9781538666326

P2P botnets represent a major threat and are responsible for a wide range of criminal activities. Fortunately, statistical algorithms that operate on NetFlow data have become a reliable means for their detection and can identify infected machines in large networks with high precision. However, many of these algorithms presume visibility on large parts of the communication within the botnet. In addition, so far only structured P2P botnets have been addressed, while especially more recent botnets shifted to unstructured P2P approaches. In this paper, we propose a detection algorithm based on the statistical concept of random walks and test it on real-world graph data of unstructured P2P botnets. Our evaluation results indicate that our algorithm yields high-precision results of about 90% at significantly reduced visibility on bots and their communication. Transferred to an Internet-wide scope, this is achieved when NetFlow data includes at least 5% of the bots and 10% of their communication links.

关键词： Botnet detection algorithms Clustering algorithms Feature extraction Support vector machines Machine learning Data mining

来源：评论

学校读者我要写书评

暂无评论

Modeling and Reconstructing Textile Sensor Noise: Implications for Wearable Technology 42

Modeling and Reconstructing Textile Sensor Noise: Implicatio...

引用

42nd Annual International Conference of the IEEE-Engineering-in-Medicine-and-Biology-Society (EMBC)

作者： Tian, Yupeng Abdizadeh, Mohammad Mahnam, Amin Bhattachan, Presish Meghrazi, Milad Alizadeh Eskandarian, Ladan Kabir, Muammar Mellal, Idir Popovic, Milos R. Lankarany, Milad Univ Toronto Inst Biomat & Biomed Engn IBBME Toronto ON Canada Univ Hlth Network Res Inst Toronto Rehabil Inst Toronto ON Canada Univ Toronto Dept Mat Sci & Engn Toronto ON Canada Myant Inc Toronto ON Canada Univ Hlth Network Krembil Res Inst Toronto ON Canada

ISBN: (数字)9781728119908

ISBN: (纸本)9781728119908

Wearable sensors enable the simultaneous recording of several electrophysiological signals from the human body in a non-invasive and continuous manner. Textile sensors are garnering substantial interest in the wearable technology because they can be knitted directly into the daily-used objects like underwear, bra, dress, etc. However, accurate processing of signals recorded by textile sensors is extremely challenging due to the very low signal-to-noise ratio (SNR). Systematic classification of textile sensor noise (TSN) is necessary to: (i) identify different types of noise and their statistical characteristics, (ii) explore how each type of noise influences the electrophysiological signal, (iii) develop optimal textile-specific electronics that suppress TSN, and (iv) reproduce TSN and create large dataset of textile sensors to validate various machine learning and signal processing algorithms. In this paper, we develop a novel technique to classify textile sensor artifacts in ECG signals. By simultaneously recording signals from the waist (textile sensors) and chest (gel electrode), we extract TSN by removing the chest ECG signal from the recorded textile data. We classify TSN based on its morphological and statistical features in two main categories, namely, slow and fast. Linear prediction coding (LPC) is utilized to model each class of TSN by auto-regression coefficients and residues. The residual signal can be approximated by Gaussian distribution which enables reproducing slow and fast artifacts that not only preserve the similar morphological features but also fulfill the statistical properties of TSN. By reproducing TSN and adding them to clean ECG signals, we create a textile-like ECG signal which can be used to develop and validate different signal processing algorithms.

关键词： Textiles Electrocardiography Feature extraction Signal to noise ratio Predictive models Signal processing algorithms detection algorithms

来源：评论

学校读者我要写书评

暂无评论

Active monitoring of the state of motion in two-wheeled vehicles in absence of a valid GPS/GNSS signal 21st

Active monitoring of the state of motion in two-wheeled vehi...

引用

21st IFAC World Congress on Automatic Control - Meeting Societal Challenges

作者： Gelmini, Simone Strada, Silvia Tanelli, Mara Savaresi, Sergio De Tommasi, Claudio Dipartimento Elettron Informaz & Bioingn Piazza Leonardo da Vinci 32 I-20133 Milan Italy AlfaEvolut Technol SpA Via Salingrado 37 I-40128 Bologna Italy

In insurance telematics, the information about the motion of a vehicle is mostly derived by a combination of inertial signals measured from e-Boxes installed integral to the vehicle. However, one must cope with poor reliability and possible lack of continuity of the GPS/GNNS signals. In this work, an inertial-based classification method that discriminates whether a two-wheeled vehicle is in motion is presented. This binary detection can be very helpful in circumstances where the GPS/GNNS signal is not sufficiently reliable and consequently the speed. With respect to what dead-reckoning algorithms do, the present contribution aims to recognize when the vehicle is moving without estimating the vehicle speed, but rather by correctly interpreting the intensity of the measured inertial signals. The approach has been extensively tested on experimental data, proving its suitability for practical applications. Copyright (C) 2020 The Authors.

关键词： Two-Wheeled Vehicles Crash detection detection algorithms Insurance Telematics Telematic e-Box

来源：评论

学校读者我要写书评

暂无评论

An acoustic emission activity detection method based on short-term waveform features: Application to metallic components under uniaxial tensile test

引用

MECHANICAL SYSTEMS AND SIGNAL PROCESSING 2020年 142卷 106753-106753页

作者： Pinal-Moctezuma, Fernando Delgado-Prieto, Miguel Romeral-Martinez, Luis Univeisitat Politecn Catalunya Dept Elect Grp Mot Control & Ind Applicat MCIA ESEIAAT Colom 11 E-08222 Terrassa Spain

The Acoustic Emission (AE) phenomenon has been used as a powerful tool with the purpose to either detect, locate or assess damage for a wide range of applications. Derived from its monitoring, one major current challenge on the analysis of the acquired signal is the proper identification and separation of each AE event. Current advanced methods for detecting events are primarily focused on identifying with high accuracy the beginning of the AE wave;however, the detection of the conclusion has been disregarded in the literature. For an automatic continuous detection of events within a data stream, this lack of accuracy for the conclusion of the events generates errors in two critical aspects. In one hand, it deteriorates the accuracy of the measurement of the events duration, truncating the span of the event, which is undesirable in evaluation applications;and in the other hand, it causes false detections. In this work, an accurate and computationally efficient AE activity detector is presented, using a framework inspired by the area of speech processing, and which provides the required indicators to accurately detect the onset and the end of an AE event. This is achieved by means of a threshold approach that instead of directly operates with the transduced voltage signal it does so over the Short-Term Energy and the Short-Term Zero-Crossing Rate measures of the signal. The STE-ZCR method is developed for an application related to the continuous monitoring of a single AE channel derived from the characterization of metallic components by means of a uniaxial tensile test. Additionally, two experimental test-benches are implemented with the aim to quantify the accuracy and the quality of event detection of the presented method. Finally, the obtained results are compared with four different techniques, representing the current state of the art related to AE activity detection. (C) 2020 Elsevier Ltd. All rights reserved.

关键词： Acoustic emission Short-time analysis Onset detection Endpoint detection detection algorithms Performance analysis

来源：评论

学校读者我要写书评

暂无评论

On the informativeness of measurements in Shiryaev's Bayesian quickest change detection

引用

AUTOMATICA 2020年 111卷 108645-000页

作者： Ford, Jason J. James, Jasmin Molloy, Timothy L. Queensland Univ Technol Sch Elect Engn & Comp Sci 2 George St Brisbane Qld 4000 Australia

This paper provides the first description of a weak practical super-martingale phenomenon that can emerge in the test statistic in Shiryaev's Bayesian quickest change detection (QCD) problem. We establish that this super-martingale phenomenon can emerge under a condition on the relative entropy between pre and post change densities when the measurements are insufficiently informative to overcome the change time's geometric prior. We illustrate this super-martingale phenomenon in a simple Bayesian QCD problem which highlights the unsuitability of Shiryaev's test statistic for detecting subtle change events. (C) 2019 Elsevier Ltd. All rights reserved.

关键词： Bayesian quickest change detection detection algorithms Markov models Super-martingale Maximal inequality

来源：评论

学校读者我要写书评

暂无评论

Design of MIMO Systems using Energy Detectors for Sub-TeraHertz Applications 31

Design of MIMO Systems using Energy Detectors for Sub-TeraHe...

引用

31st Annual IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (IEEE PIMRC)

作者： Bicais, Simon Dore, Jean-Baptiste Savin, Valentin CEA Leti Minatec Campus Grenoble France

ISBN: (纸本)9781728144900

The significant amount of unused spectrum in sub-TeraHertz frequencies is contemplated to realize high rate wireless communications for beyond 5G networks. Yet, the performance of sub-TeraHertz systems is severely degraded by strong oscillator phase noise. Therefore, we investigate in this paper the use of multiple-input multiple-output (MIMO) systems with energy detection receivers to achieve high rate communications robust to phase noise. First, the design of the receiver detection algorithm is addressed. We derive the joint maximum likelihood detector corresponding to the studied nonlinear MIMO channel. Second, the system performance is assessed through numerical simulations. We consider a realistic scenario modeling an indoor wireless link with directive antennas and strongly correlated line-of-sight channels. Our results demonstrate that high rate sub-TeraHertz systems can be implemented on low-complexity and low-power architectures using MIMO systems and energy detection receivers.

关键词： Sub-TeraHertz communications MIMO Phase noise Envelope detectors detection algorithms

来源：评论

学校读者我要写书评

暂无评论

Overlapping Community detection Method Based on Network Representation Learning and Density Peaks

引用

IEEE ACCESS 2020年 8卷 226506-226514页

作者： Liu, Hongtao Li, Gege Chongqing Univ Posts & Telecommun Sch Comp Sci & Technol Chongqing 400000 Peoples R China

At present, the research on complex social networks has attracted extensive attention from scholars, and community detection is an important research direction in the study of network structure. Network data is often high-dimensional and very large, which makes it very difficult to process. Therefore, it is of great significance for community detection to represent network structure with low-dimensional vector. And many real world social networks contain overlapping communities. In this paper, we propose an overlapping community detection method based on network representation learning and density peaks, called NRLDP. First, it uses network representation learning technology to represent the unweighted network or weighted network with low-dimensional vectors. Then, it applies the density peaks clustering algorithm to overlapping community detection, uses cosine similarity to calculate the distance between nodes, and improves the local density calculation method. Finally, it selects the core node according to the relative distance and local density, and allocates the remaining nodes to achieve overlapping community detection of unweighted network or weighted network. Compared with relevant community detection methods on real world social networks and synthetic networks of LFR Benchmark, the results of the experiment show that our proposed approach is effective and accurate.

关键词： Clustering algorithms Social networking (online) detection algorithms Partitioning algorithms Optimization Licenses Complex networks Network representation learning density peaks overlapping community detection

来源：评论

学校读者我要写书评

暂无评论

Cloud to Cable: A Drone Framework for Autonomous Power line Inspection 23

Cloud to Cable: A Drone Framework for Autonomous Power line ...

引用

23rd Euromicro Conference on Digital System Design (DSD)

作者： Schofield, Oscar Bowen Lorenzen, Kasper Hoj Ebeid, Emad Univ Southern Denmark SDU UAS Ctr Maersk Mc Kinney Moller Inst Odense Denmark

ISBN: (数字)9781728195353

ISBN: (纸本)9781728195353

A large percentage of current overhead power transmission infrastructure is becoming aged, raising the demand for frequent grid inspections in an efficient, and cost-effect manner in order to keep it functional. Autonomous drones are a promising solution for infrastructure inspection, but require further investigation and development to be a viable alternative. This paper proposes a drone framework based on open-source platforms for autonomous inspection of electrical grid infrastructure and for cable grasping. The framework incorporates a cloud service functionality to provide geo-location data of nearby power pylons;a localisation system for navigation near power pylons and grasping cables;and algorithms for finding and planning paths for navigation and grasping over cables. The proposed framework has been tested and validated in simulation, with the development of a UAV platform for future integration.

关键词： Cloud services detection algorithms Autonomous missions Drones Inspection Framework ROS

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：