检索结果-内蒙古大学图书馆

IEEE International Symposium on Circuits and Systems (ISCAS)

作者： Xie, Chen Daghero, Francesco Chen, Yukai Castellano, Marco Gandolfi, Luca Calimera, Andrea Macii, Enrico Poncino, Massimo Pagliari, Daniele Jahier Politecn Torino Turin Italy STMicroelectronics Cornaredo Italy

ISBN: (纸本)9781665484855

Low-resolution infrared (IR) array sensors offer a low-cost, low-power, and privacy-preserving alternative to optical cameras and smartphones/wearables for social distance monitoring in indoor spaces, permitting the recognition of basic shapes, without revealing the personal details of individuals. In this work, we demonstrate that an accurate detection of social distance violations can be achieved processing the raw output of a 8x8 IR array sensor with a small-sized Convolutional neural network (CNN). Furthermore, the CNN can be executed directly on a Microcontroller (MCU)-based sensor node. With results on a newly collected open dataset, we show that our best CNN achieves 86.3% balanced accuracy, significantly outperforming the 61% achieved by a state-of-the-art deterministic algorithm. Changing the architectural parameters of the CNN, we obtain a rich Pareto set of models, spanning 70.5-86.3% accuracy and 0.18-75k parameters. Deployed on a STM32L476RG MCU, these models have a latency of 0.73-5.33ms, with an energy consumption per inference of 9.38-68.57 mu J.

关键词： Social Distance COVID-19 Edge computing Infrared Sensors Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

7th International Conference on Soft computing in Data Science, SCDS 2023

7th International Conference on Soft Computing in Data Scien...

引用

7th International Conference on Soft computing in Data Science, SCDS 2023

ISBN: (纸本)9789819904044

The proceedings contain 21 papers. The special focus in this conference is on Soft computing in Data Science. The topics include: Short-Time Fourier Transform with Optimum Window Type and Length: An Application for Sag, Swell and Transient;cox Point Process with Ridge Regularization: A Better Approach for Statistical Modeling of Earthquake Occurrences;discovering Popular Topics of Sarawak Gazette (SaGa) from Twitter Using Deep Learning;Comparison Analysis of LSTM and CNN Variants with Embedding Word Methods for Sentiment Analysis on Food Consumption Behavior;towards Robust Underwater Image Enhancement;a Comparative Study of Machine Learning Classification models on Customer Behavior Data;performance Evaluation of Deep Learning Algorithms for Young and Mature Oil Palm Tree Detection;object Detection Based Automated optical Inspection of Printed Circuit Board Assembly Using Deep Learning;the Characterization of Rainfall Data Set Using Persistence Diagram and Its Relation to Extreme Events: Case Study of Three Locations in Kemaman, Terengganu;fault Diagnosis Methods of Deep Convolutional Dynamic Adversarial networks;Clustering Stock Prices of Industrial and Consumer Sector Companies in Indonesia Using Fuzzy C-Means and Fuzzy C-Medoids Involving ACF and PACF;zero-Inflated Time Series Model for Covid-19 Deaths in Kelantan Malaysia;carbon-Energy Composite Flow for Transferred Multi-searcher Q-Learning Algorithm with Reactive Power Optimization;multi-source Heterogeneous Data Fusion Algorithm Based on Federated Learning;dynamic Micro-cluster-Based Streaming Data Clustering Method for Anomaly Detection;federated Ensemble Algorithm Based on Deep neural network;performance Comparison of Feature Selection Methods for Prediction in Medical Data.

关键词：

来源：评论

学校读者我要写书评

暂无评论

MacularNet: Towards Fully Automated Attention-Based Deep CNN for Macular Disease Classification

引用

SN Computer Science 2022年第2期3卷 142页

作者： Mishra, Sapna S. Mandal, Bappaditya Puhan, Niladri B. School of Electrical Sciences Indian Institute of Technology Bhubaneswar India School of Computing and Mathematics Keele University Newcastle United Kingdom

In this work, we propose an attention-based deep convolutional neural network (CNN) model as an assistive computer-aided tool to classify common types of macular diseases: age-related macular degeneration, diabetic macular edema, diabetic retinopathy, choroidal neovascularization, macular hole, and central serous retinopathy from normal macular conditions with the help of scans from optical coherence tomography (OCT) imaging. Our proposed architecture unifies refined deep pre-trained models using transfer learning with limited training data and a deformation-aware attention mechanism encoding crucial morphological variations appearing in the deformation of retinal layers, detachments from the subsequent layers, presence of fluid-filled regions, geographic atrophy, scars, cysts, drusen, to achieve superior macular imaging classification performance. The proposed attention module facilitates the base network to automatically focus on the salient features arising due to the macular structural abnormalities while suppressing the irrelevant (or no cues) regions. The superiority of our proposed method lies in the fact that it does not require any pre-processing steps such as retinal flattening, denoising, and selection of a region of interest making it fully automatic and end-to-end trainable. Additionally, it requires a reduced number of network model parameters while achieving higher diagnostic performance. Extensive experimental results, analysis on four datasets along with the ablation studies show that the proposed architecture achieves state-of-the-art performance. © The Author(s) 2022.

来源：评论

学校读者我要写书评

暂无评论

Handwritten Hindko Text Recognition Using Deep neural network

Handwritten Hindko Text Recognition Using Deep Neural Networ...

引用

Frontiers of Information Technology (FIT)

作者： Tanveer Ahmed Sohail Khan Khalil Khan School of Computing Sciences Pak-Austria Fachhochschule Haripur Pakistan Sino-Pak Centre for Artificial Intelligence Pak-Austria Fachhochschule Haripur Pakistan Department of Computer Science Nazarbayev University Kazakhstan

ISBN: (数字)9798331510503

ISBN: (纸本)9798331510510

Handwritten text recognition is a challenging task with significant implications for cultural preservation and technological advancement. In this research paper, we propose a unique approach for Hindko Handwritten text recognition. The primary objectives of the study were to develop an accurate and efficient system capable of recognizing Hindko handwritten text and to compare the performance of deep learning models, Convolutional neural networks (CNN) and Recurrent neural networks (RNN) and HTR-Attention model. To achieve these objectives, we collected a substantial dataset of handwritten Hindko text, comprising diverse writing styles and variations. Processing techniques were applied to enhance the data quality and extract meaningful features for training the deep neural networks. Through rigorous experimentation and evaluation, the research findings demonstrated promising results. The HTR-Attention model achieved and impressive accuracy of 88.97%. while CNN and RNN models attained an accuracy of 88.36% and 80.45% respectively. The potential applications of this research encompass various fields, including language recognition systems, optical character recognition systems, and language translation technologies. Furthermore, the developed model can be integrated into digital learning platforms to facilitate Hindko language learners and preserve the cultural heritage associated with this regional Language.

关键词： Measurement Deep learning Accuracy Recurrent neural networks Text recognition optical character recognition Feature extraction Cultural differences Convolutional neural networks Character recognition

来源：评论

学校读者我要写书评

暂无评论

Nonlinear optical feature generator for machine learning

引用

APL PHOTONICS 2023年第10期8卷

作者： Yildirim, Mustafa Oguz, Ilker Kaufmann, Fabian Escale, Marc Reig Grange, Rachel Psaltis, Demetri Moser, Christophe Ecole Polytech Fed Lausanne EPFL Lab Appl Photon Devices Lausanne Switzerland Swiss Fed Inst Technol Inst Quantum Elect Dept Phys Opt Nanomat Grp Zurich Switzerland Versics AG Auguste Piccard Hof 1HPT Bldg CH-8093 Zurich Switzerland Ecole Polytech Fed Lausanne EPFL Opt Lab Lausanne Switzerland

Modern machine learning models use an ever-increasing number of parameters to train (175 x 10(9) parameters for GPT-3) with large datasets to achieve better performance. optical computing has been rediscovered as a potential solution for large-scale data processing, taking advantage of linear optical accelerators that perform operations at lower power consumption. However, to achieve efficient computing with light, it remains a challenge to create and control nonlinearity optically rather than electronically. In this study, a reservoir computing approach (RC) is investigated using a 14-mm waveguide in LiNbO3 on an insulator as an optical processor to validate the benefit of optical nonlinearity. Data are encoded on the spectrum of a femtosecond pulse, which is launched into the waveguide. The output of the waveguide is a nonlinear transform of the input, enabled by optical nonlinearities. We show experimentally that a simple digital linear classifier using the output spectrum of the waveguide increases the classification accuracy of several databases by similar to 10% compared to untransformed data. In comparison, a digital neural network (NN) with tens of thousands of parameters was required to achieve similar accuracy. With the ability to reduce the number of parameters by a factor of at least 20, an integrated optical RC approach can attain a performance on a par with a digital NN.

关键词： Waveguides

来源：评论

学校读者我要写书评

暂无评论

A fast learning approach for autonomous navigation using a deep reinforcement learning method

引用

ELECTRONICS LETTERS 2021年第2期57卷 50-53页

作者： Ejaz, Muhammad Mudassir Tang, Tong Boon Lu, Cheng-Kai Univ Teknol Petronas Dept Elect & Elect Engn Seri Iskandar Perak Malaysia Univ Teknol Petronas Inst Hlth & Analyt Seri Iskandar Perak Malaysia

Deep reinforcement learning-based methods employ an ample amount of computational power that affects the learning process. This paper proposes a novel approach to speed up the training process and improve the performance of autonomous navigation for a tracked robot. The proposed model named "layer normalization dueling double deep Q-network" has been trained in a virtual environment and then implemented it to a tracked robot for testing in a real-world scenario. Depth images have been used instead of RGB images to preserve the temporal information. Features are extracted using convolutional neural networks, and actions are derived using the dueling double deep Q-network. The input data has been normalized before each convolutional layer, which reduces the covariate shift by 69%. This end-to-end network architecture of the proposed model provides stability to the network, relieves the burden of computational cost, and converges in much less number of episodes. Compared with three Q-variant models, the proposed model demonstrates outstanding performance in terms of episodic reward and convergence rate. The proposed model took 12.8% fewer episodes for training compared to other models.

关键词： optical, image and video signal processing Robotics Computer vision and image processing techniques Control engineering computing Reinforcement learning neural nets

来源：评论

学校读者我要写书评

暂无评论

基于Neuroptica的光学神经网络仿真实验设计

引用

实验室研究与探索 2023年第11期42卷 135-140页

作者：卢瑾苏豪宇王艳任宏亮浙江工业大学计算机科学与技术学院杭州310023 浙江工业大学信息工程学院杭州310023

光学神经网络光电子技术与神经网络模型,有望突破传统电子神经网络计算速度和网络功率效率低的技术瓶颈。为解决教学活动中缺乏实际演示案例问题,设计了利用光学神经网络进行手写数字识别实验。利用光学神经网络芯片级仿真平台Neuroptic... 详细信息

光学神经网络光电子技术与神经网络模型,有望突破传统电子神经网络计算速度和网络功率效率低的技术瓶颈。为解决教学活动中缺乏实际演示案例问题,设计了利用光学神经网络进行手写数字识别实验。利用光学神经网络芯片级仿真平台Neuroptica构建手写数字识别系统,学生在不同参数下训练模型进行对比实验,也可自主编程设计ONN模型实现其他机器学习任务以验证模型的优劣。

关键词：光电子技术光学神经网络马赫曾德干涉仪手写数字识别

来源：评论

学校读者我要写书评

暂无评论

Automated identification of ocular toxoplasmosis in fundoscopic images utilizing deep learning models

Automated identification of ocular toxoplasmosis in fundosco...

引用

International Conference on Computer and Information Technology (ICCIT)

作者： Arsi Zaman Prionto Kumar Choudhury Rizvee Rifat Chowdhury Asma Akter Anika Sumaiya Rahman Ramisa BRAC University Dhaka Bangladesh

ISBN: (数字)9798331519094

ISBN: (纸本)9798331519100

The detection of ocular toxoplasmosis is based on eye fundus images, which are produced and analysed by a specialist optician. Although Deep Learning techniques are being quickly adopted in many areas, their application in ocular diagnosis has not been widely studied. In this research, we created a highly effective Convolutional neural network (CNN) model that can accurately identify and classify ocular Toxoplasmosis (OT) images into four categories: healthy, active, inactive, and active-inactive. After further examination, we merged three of these categories—active, inactive, and active-inactive—into a single class called "unhealthy," resulting in a binary classification of healthy and unhealthy images. Our ensemble model shows 96% accuracy. Our CNN model shows a high level of accuracy in distinguishing between these two categories. To validate the performance of our custom CNN model, we compared it with three pre-trained models (VGG16, VGG19, and MobileNet) using the same dataset. The findings revealed that both our proposed CNN model and the pre-trained architectures displayed comparable performance metrics, including accuracy, recall, F1 score, and precision. Our model achieved a remarkable accuracy of 97%, surpassing the performance of previously utilised models in diagnosing retinal disorders.

关键词： Deep learning Accuracy neural networks optical computing Predictive models Retina optical imaging Convolutional neural networks Information technology Diseases

来源：评论

学校读者我要写书评

暂无评论

OPIMA: optical Processing-In-Memory for Convolutional neural network Acceleration

arXiv

引用

arXiv 2024年

作者： Sunny, Febin Shafiee, Amin Balasubramaniam, Abhishek Nikdast, Mahdi Pasricha, Sudeep Electrical and Computer Engineering Department Colorado State University Fort CollinsCO United States

Recent advances in machine learning (ML) have spotlighted the pressing need for computing architectures that bridge the gap between memory bandwidth and processing power. The advent of deep neural networks has pushed traditional Von Neumann architectures to their limits due to the high latency and energy consumption costs associated with data movement between the processor and memory for these workloads. One of the solutions to overcome this bottleneck is to perform computation within the main memory through processing-in-memory (PIM), thereby limiting data movement and the costs associated with it. However, DRAM-based PIM struggles to achieve high throughput and energy efficiency due to internal data movement bottlenecks and the need for frequent refresh operations. In this work, we introduce OPIMA, a PIM-based ML accelerator, architected within an optical main memory. OPIMA has been designed to leverage the inherent massive parallelism within main memory while performing high-speed, low-energy optical computation to accelerate ML models based on convolutional neural networks. We present a comprehensive analysis of OPIMA to guide design choices and operational mechanisms. Additionally, we evaluate the performance and energy consumption of OPIMA, comparing it with conventional electronic computing systems and emerging photonic PIM architectures. The experimental results show that OPIMA can achieve 2.98× higher throughput and 137× better energy efficiency than the best-known prior work. © 2024, CC BY-NC-ND.

关键词： Silicon photonics

来源：评论

学校读者我要写书评

暂无评论

Analog Photonics computing for Information Processing, Inference and Optimisation

arXiv

引用

arXiv 2023年

作者： Stroev, Nikita Berloff, Natalia G. Department of Physics of Complex Systems Weizmann Institute of Science Rehovot76100 Israel Department of Applied Mathematics and Theoretical Physics University of Cambridge CambridgeCB3 0WA United Kingdom

This review presents an overview of the current state-of-the-art in photonics computing, which leverages photons, photons coupled with matter, and optics-related technologies for effective and efficient computational purposes. It covers the history and development of photonics computing and modern analogue computing platforms and architectures, focusing on optimization tasks and neural network implementations. The authors examine special-purpose optimizers, mathematical descriptions of photonics optimizers, and their various interconnections. Disparate applications are discussed, including direct encoding, logistics, finance, phase retrieval, machine learning, neural networks, probabilistic graphical models, and image processing, among many others. The main directions of technological advancement and associated challenges in photonics computing are explored, along with an assessment of its efficiency. Finally, the paper discusses prospects and the field of optical quantum computing, providing insights into the potential applications of this technology. © 2023, CC BY.

关键词： optical data processing

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：