检索结果-内蒙古大学图书馆

Proceedings of the 11th International conference on Distributed Smart Cameras

作者： Walther Carballo-Hernández Miguel Arias-Estrada Instituto Nacional de Astrofísica Óptica y Electrónica Computer Science Department Tonanzintla Puebla Mexico

ISBN: (纸本)9781450354875

The biological human brain works with spiking neural networks which computational complexity is simple, compensated by the high density connection between neurons. However, most of our research in artificial neural networks is based on simplified models that need many real values processing elements of complex computing which require too much silicon space, energy and slow learning convergence. Therefore, in this brief article, it is presented a proposal for a fully digital architecture on FPGAs optimized for a highly dense intersynaptic connection on an event-based quantized Sigma-Delta pulse coding for deep Convolutional Neural Networks. This article presents early results of an approach of neuromorphic hardware design for information or pixel luminosity changes coded in time, using Sigma-Delta modulation, the design of a Pulsed Arithmetic-Logic Unit for bitstream operations with quantized weights reducing memory, from 32 bits in floating point representation, down to 1 bit. This 32x memory reduction and binary operation cells in a systolic architectures makes the integration of deep learning models feasible for embedded design like smart cameras using FPGAs.

关键词： Artificial Neural Networks (ANN) Systolic arrays Deep Learning Pulsed Multiplier Field Programmable Gate Array (FPGA) Synaptic weight Activation function Feature map Pulsed Arithmetic Logic Unit Convolutional Neural Networks (CNN) Sigma-Delta (SD) modulation Digital signal processing Pulsed Adder Event-based Neuromorphic Systems

来源：评论

学校读者我要写书评

暂无评论

Artificial Intelligence to Impart Surveillance, Tracking, & Actuation on Suspicious Activities

Artificial Intelligence to Impart Surveillance, Tracking, & ...

引用

IEEE International Advance Computing conference, IACC

作者： Surajit Bhattacherjee G.C. Somashekhar Department of Electronics & Communication Rajiv Gandhi Institute of Technology Bangalore Bangalore India

ISBN: (纸本)9781509015610

Along with the fast growing economy of a developing country as far as India is concerned, there has been reported an emerging number of crime and criminal offences. In many of the cases the suspects are quite freely escaped because of the insufficient testimonials and lack of communication links in time. As an instance, suppose a bank robbery takes place and right then no one is able to press the emergency buzzer or to dial up to nearest police station. Eventually, when the nearest police station is informed about the fact, the gang is beyond the reach leaving no clue behind. Same is observed in many cases as rash driving and hit and run, eve teasing and molestation of women at less crowded streets, pick pocketing and blackmailing at Automatic Teller Machines at roadside and many more. Where there is a crime involved, there has to be a quality surveillance and expert actuation to stand against all these activities with minimum human intervention. Our goal is to design an intelligent device that can continuously keep an eye on any activity happening around it and on administrator's request it can track any suspect, be it an object or a human being, with minimum power consumption and least human efforts. As an actuator it sends a signal to the nearest traffic control or police station intimating the state of crime along with its location and also tells about which direction it is heading to. In view of a real-time establishment of the concept, the flying device is designed to be a robot that looks and flies like a Bird, the processor is chosen to be ARDUINO UNO, the image processing is made by mat-lab tool and the communication link between the administrator and the device is via RFID system and between device and the police control is achieved by GPS/GSM technology.

关键词： Law enforcement Birds Automobiles Traffic control image coding Surveillance Robots

来源：评论

学校读者我要写书评

暂无评论

Nanopositioning MEMS stage for high speed positioning of Metamaterials lenses for use in high resolution optical imaging

Nanopositioning MEMS stage for high speed positioning of Met...

引用

IEEE International conference on Advanced Intelligent Mechatronics (AIM)

作者： Herfst, R. W. Bijster, R. J. F. Dekker, A. Wei, J. van Zeijl, H. W. Kruidhof, R. Sadeghian, H. TNO Tech Sci Optomechatron Dept Stieltjesweg 1 NL-2628 CK Delft Netherlands Else Kooi Lab Feldmannweg 17 NL-2628 CT Delft Netherlands

ISBN: (纸本)9781509060009

High resolution and high throughput imaging are typically mutually exclusive. While there is a wide range of techniques to image features beyond the diffraction limit of light, they all have their own benefits and drawbacks, but they are often very slow compared to optical systems. As such, extending the performance of optical microscopes remains. Examples of high resolution optical concepts are metamaterials nano-antennas, superoscillatory lenses and hyperlenses. These concepts require precise and high speed positioning of the optical element at lens-to-sample separations measured in tens of nanometers. We present a design and process for a MEMS nanopositioning stage capable of sub-nm positioning of a metamaterials optical lens with very high bandwidth. Laser vibrometer measurements on the first batch of fully fabricated devices showed that the first eigenfrequency is at 500 +/- 25 kHz, somewhat below the 660 kHz obtained by FEM simulations. Due to squeeze film damping, the quality factor is relatively low (approximately 2-2.5), which is advantageous for reaching high positioning bandwidth when the device is used in a near-field optical imaging microscope.

关键词： Estimation Field programmable gate arrays OFDM Modulation Filtration Clocks signal processing algorithms

来源：评论

学校读者我要写书评

暂无评论

Individual Prediction of Brain Tumor Histological Grading Using Radiomics on Structural MRI

Individual Prediction of Brain Tumor Histological Grading Us...

引用

IEEE Symposium on Nuclear Science (NSS/MIC)

作者： Stijn Bonte Ingeborg Goethals Roel Van Holen Department of Electronics and Information Systems Medical Image and Signal Processing (MEDISIP) and the Department of Nuclear Medicine Ghent University Ghent Belgium

The accurate diagnosis of brain tumors is of primary importance for optimal therapy planning. In clinical practice, this is determined on a biopsy, exposing the patient to the risk of complications. Moreover, sampling bias and performer variability may influence the result. Several studies have investigated the histological grading of brain tumors in a non-invasive way by extracting features from medical images. A multicenter study where both tumor grade and cell type are simultaneously predicted is however lacking. In this study we collected structural MRI-scans from 294 patients with glioma acquired in different centers (the local hospital and two online databases). The goal was to predict tumor grade and cell type of individual patients using a radiomics study with Random Forests. In a multiclass design, we obtain a global accuracy of 59.9% to predict tumor grade and 53.4% to predict cell type. Converting the problem to binary classification, we obtain an accuracy of 98.3% to distinguish between meningioma and glioma, and 84.5% to distinguish between low-grade glioma and glioblastoma. A high degree of diagnosis variability and overlap between different low-grade classes might cause the reduced prediction accuracy. Our results however show that radiomics on structural MRI is a suitable approach for non-invasively assessing brain tumor diagnosis and might be used for individual treatment planning.

关键词： Tumors Feature extraction Magnetic resonance imaging Medical treatment Biomedical imaging Forestry

来源：评论

学校读者我要写书评

暂无评论

CS-MRA reconstruction based on nonsubsampled contourlet transform in frequency domain (NSCT-FD)

CS-MRA reconstruction based on nonsubsampled contourlet tran...

引用

International conference on Systems and Informatics (ICSAI)

作者： Yang Heng Xiu-mei Chen Wei Wang Min Tang School of Electronics and Information Nantong University Nantong China

Compressed sensing (CS) is a novel developed theoretical framework for information acquisition and processing. Taking advantages of the sparsity or compressibility inherent in real world signals, CS can collect compressed data at the sampling rate much lower than that needed in Shannon's theorem. When it has been used for medical imaging techniques, CS theory can fast the scanning speed of MRI, reduce the radiation doses and alleviate the patients' suffering. In order to obtain the effective sparse representation, the nonsubsampled contourlet transform in the frequency domain (NSCT-FD) is adopted, then an improved fast iterative soft thresholding algorithm (FISTA) is applied for CS-MRA reconstruction. This novel method not only inherits the simplicity and effectiveness of the original FISTA, and the sparse curvy representation ability of the contourlet transform, but also has the advantages of the shift-invariant property and much less computational burden. The performance is evaluated qualitatively and quantitatively both in noiseless and noisy situations, compared to the classic wavelet and the sharp frequency localization contourlet transform (SFLCT) method. Three quantitative indices are employed including the peak signal to noise ratio (PSNR), mutual information (MI) and relative l 2 norm error (RLNE) and qualitative performance evaluations use the profiles in horizontal direction and local region magnification comparison. Experimental results demonstrate the superiority of the NSCT-FD algorithm, displaying higher PSNR and MI, and lower RLNE indices in both noiseless and noisy MRA images, which have good reconstruction accuracy with reasonable real time computation speed. It is manifest that the NSCT-FD algorithm can be applied in fast medical imaging fields to achieve quality images in a relatively fast process, which will benefit cardiac and carotid MRI, dynamic MRI and MRA.

关键词： Transforms image reconstruction Filter banks Frequency-domain analysis Algorithm design and analysis Filtering algorithms

来源：评论

学校读者我要写书评

暂无评论

A Digital Control Mechanism for the Delay of a Dual-Microphone Analog Beamformer

A Digital Control Mechanism for the Delay of a Dual-Micropho...

引用

International conference on Optimization of Electrical and Electronic Equipment

作者： Paul Faragó Sorin Hintea Florin Sandu Technical University of Cluj-Napoca Transilvania University of Brasov

ISBN: (纸本)9781509044900

The microelectronic industry is driven by the continuous demand for processing capabilities and performance. As such, most processing is performed nowadays in a digital fashion. Analog however is not to be neglected. A key feature in modern mixed-signal processing is wide-range tuning of the analog processing sections. One step forward is the implementation of digitally programmable analog. The aim of this paper is to illustrate the digital tuning mechanism on the analog beamformer of a dual-microphone array. Analog processing consists of having one microphone signal delayed by an allpass filter, in order to be synchronized to the other microphone signal. A Gm-C filter implementation was considered, and filter tuning was performed via the transconductor bias current. Digital control of the bias current was implemented with a current digital to analog converter, enabling a straightforward and simple control mechanism. Transistor-level simulation results validate the proposed design.

关键词： signal and image processing filtering

来源：评论

学校读者我要写书评

暂无评论

Real Time SoC architectures for Analysis of EEG

Real Time SoC architectures for Analysis of EEG

引用

International conference on Computing, Communication, Control and Automation

作者： Varsha K. Harpale Vinayak K. Bairagi S. S. Pujari AISSMS PCCOE Pune-44 E&TC AISSMS IOIT Pune-01 E&TC Department PCCOE Pune-44

Electroencephalography (EEG) is a multichannel non-stationary brain signal composed of multi-frequency. EEG signals are recorded with multiple electrodes either by scalp or intracranial method. It is well proven tool for brain disease detection or prediction. These brain signals are recorded with multiple electrodes with sampling rate. Thus the very huge data need to process and hence computationally intensive. Different architectures have been proposed and implemented to get high performance with accuracy from the device. This proposed work covers comparative study of different efficient architectures of the EEG signal analyzer and designs biosignal processor using FPGA by establishing link between FPGA and PC with URART serial link. More or less of the architectures of EEG signal processing are mainly used for seizure detection. The basic objective of this paper is to understand the implementation of real tiem EEG analyser using FPGA and SoC, to design simple bisignal processor for EEG analysis and analyze architectures based on different performance parameters like frequency of operation, peak power consumption, technology, memory size (bits), accuracy, ADC Resolution, chip area, gates/SRAM.

关键词： Electroencephalogram (EEG) Seizures System on Chip (SoC) Independent Component Analysis (ICA)

来源：评论

学校读者我要写书评

暂无评论

A novel multimodal fusion framework for early diagnosis and accurate classification of COVID-19 patients using X-ray images and speech signal processing techniques

引用

COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE 2022年 226卷 107109-107109页

作者： Kumar, Santosh Chaube, Mithilesh Kumar Alsamhi, Saeed Hamood Gupta, Sachin Kumar Guizani, Mohsen Gravina, Raffaele Fortino, Giancarlo Int Inst Informat Technol Dept Comp Sci & Engn Naya Raipur Chhattishgarh India Int Inst Informat Technol Dept Math Sci Naya Raipur Chhattishgarh India Natl Univ Ireland Insight Ctr Data Analyt Galway Ireland IBB Univ Fac Engn Ibb Yemen Shri Mata Vaishno Devi Univ Sch Elect & Commun Engn Katra India Mohamed Bin Zayed Univ Artificial Intelligence Machine Learning Dept Abu Dhabi U Arab Emirates Univ Calabria Dept Informat Modeling Elect & Syst Engn I-87036 Arcavacata Di Rende Italy

Background and objective: COVID-19 outbreak has become one of the most challenging problems for human being. It is a communicable disease caused by a new coronavirus strain, which infected over 375 million people already and caused almost 6 million deaths. This paper aims to develop and design a framework for early diagnosis and fast classification of COVID-19 symptoms using multimodal Deep Learning ***: we collected chest X-ray and cough sample data from open source datasets, Cohen and datasets and local hospitals. The features are extracted from the chest X-ray images are extracted from chest X-ray datasets. We also used cough audio datasets from Coswara project and local hospitals. The publicly avail-able Coughvid DetectNow and Virufy datasets are used to evaluate COVID-19 detection based on speech sounds, respiratory, and cough. The collected audio data comprises slow and fast breathing, shallow and deep coughing, spoken digits, and phonation of sustained vowels. Gender, geographical location, age, pre-existing medical conditions, and current health status (COVID-19 and Non-COVID-19) are ***: The proposed framework uses the selection algorithm of the pre-trained network to determine the best fusion model characterized by the pre-trained chest X-ray and cough models. Third, deep chest X-ray fusion by discriminant correlation analysis is used to fuse discriminatory features from the two models. The proposed framework achieved recognition accuracy, specificity, and sensitivity of 98.91%, 96.25%, and 97.69%, respectively. With the fusion method we obtained 94.99% ***: This paper examines the effectiveness of well-known ML architectures on a joint collection of chest-X-rays and cough samples for early classification of COVID-19. It shows that existing methods can effectively used for diagnosis and suggesting that the fusion learning paradigm could be a crucial asset in diagnosing future unknown illnesses. The proposed frame

关键词： Deep learning COVID-19 Early detection Speech processing X-ray image classification Multimodel fusion

来源：评论

学校读者我要写书评

暂无评论

SAR image change detection algorithm based on stationary wavelet and bi-dimensional intrinsic mode function

引用

IOP conference Series: Materials Science and Engineering 2017年第1期231卷

作者： S.Q. Huang Z.L. Wang T.G. Xie Z.C. Li School of Information Engineering Xijing University Xi'an 710123 China

Speckle noise in synthetic aperture radar (SAR) image is produced by the coherent imaging mechanism, which brings a great impact on the change information acquisition of multi-temporal SAR images. Two-dimensional stationary wavelet transform (SWT) and bi-dimensional empirical mode decomposition (BEMD) are the non-stationary signal processing theory of multi-scale transform. According to their implementation process and SAR image characteristic, this paper proposed a new multi-temporal SAR image change detection method based on the combination of the stationary wavelet transform and the bi-dimensional intrinsic mode function (BIMF) features, called SWT-BIMF algorithm. The contribution of the new algorithm includes two aspects. One is the design of the two selections of decomposition features, that is, the speckle noise filtering; another is the selected features to perform the enhance processing, so more effective change information will obtain. The feasibility of the SWT-BIMF algorithm is verified by the measured SAR image data, and good experimental results are obtained.

关键词：

来源：评论

学校读者我要写书评

暂无评论

GAMM Annual Meeting – Weimar and Ilmenau 2017, Overview of the Sections

引用

PAMM 2018年第1期17卷

Volume 17 (2017) of PAMM “Proceedings in Applied Mathematics and Mechanics” assembles the contributions to the 88 th Annual Meeting of the Gesellschaft für Angewandte Mathematik und Mechanik (GAMM), held 6 – 10 March 2017 at Weimar, Germany, supported by Bauhaus-Universität Weimar and Technische Universität Ilmenau. The contributions are grouped according to the minisymposia and sessions of the conference. Overview of the Sections Plenary Lectures PL Mechanics GAMM related DFG Priority Programmes PP01 Reliable simulation techniques in solid mechanics. Development of non-standard discretization methods, mechanical and mathematical analysis PP02 Reactive bubbly flows PP04 Calm, smooth and smart PP05 Non-smooth and complementarity-based distributed parameter systems: simulation and hierarchical optimization PP06 Polymorphic uncertainty modelling for the numerical design of structures PP07 Turbulent superstructures Minisymposia M 1 Innovative discretization methods, mechanical and mathematical investigations M 2 Recent trends in phase-field modelling M 3 Dislocation-based plasticity: State of the art and challenges M 6 Turbulent liquid metal and magnetohydrodynamic flows M 7 Flow Separation and vortical phenomena: Simulation in progress Sections 1–23 1 Multi-body dynamics 2 Biomechanics 3 Damage and fracture mechanics 4 Structural mechanics 5 Nonlinear oscillations 6 Material modelling in solid mechanics 7 Coupled problems 8 Multiscales and homogenization 9 Laminar flows and transition 10 Turbulence and reactive flows 11 Interfacial flows 12 Waves and acoustics 13 Flow control 14 Applied analysis 15 Uncertainty quantification 16 Optimization 17 Applied and numerical linear algebra 18 Numerical methods of differential equations 19 Optimization of differential equations 20 Dynamics and control 21 Mathematical signal and image processing 22 Scientific computing 23 Applied operator theory Young Researchers' Minisymposia YR 2 Computational techniques for bayesian inverse pro

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：