检索结果-内蒙古大学图书馆

Physical Assurance and Inspection of Electronics (PAINE), 2020 IEEE

作者： Md Alimoor Reza Manjulata Garimella Sami Alajouni Kerry Maize Peter Bermel David Crandall Ali Shakouri Drake University Des Moines Iowa Indiana University Bloomington Indiana Purdue University West Lafayette Indiana

Understanding and predicting microelectronic failures is important for ensuring the reliability of modern electronic devices. In this paper, we develop a set of computer vision algorithms for modeling device failure due to electromigration, in which metal atoms are displaced due to current flow. The experimental setup contains a series of optical and thermal images of aluminum interconnects. We propose deep neural networks for two specific problems: predicting where in the device a failure will occur, and estimating the remaining lifespan before failure. We pose the former as a segmentation problem and solve it with a convolutional neural network (CNN) trained on multi-scale optical images. We pose the latter as a regression problem for which we designed a convolutional neural network augmented with a recurrent module (RNN) to model the temporal dimension. We compare against two baseline networks, finding that our model can predict the age of the aluminum more accurately from the optical images instead of the thermal ones. This work is the first that pursues a deep neural network-based modeling to predict electromigration failures from experimental optical and thermal images.

关键词： Electromigration optical interconnections Aluminum optical computing optical fiber networks Predictive models optical imaging

来源：评论

学校读者我要写书评

暂无评论

Retinal Disease for Clasification Multilabel with Applying Convolutional neural networks Based Support Vector Machine and DenseNet

Retinal Disease for Clasification Multilabel with Applying C...

引用

International Seminar on Research of Information Technology and Intelligent Systems (ISRITI)

作者： Alicia Anggelia Lumbantoruan Alhadi Bustamam Prasnurzaki Anki Faculty of Mathematics and Natural Science University of Indonesia Depok Indonesia

ISBN: (纸本)9781665401524

The retina is an essential part of the eye and works to transmit visual information to the brain. In maintaining the eye, an ophthalmologist needs regular examinations, but the price is expensive and takes time. Therefore, technological developments are expected to help the medical world to detect diseases. The technology is image processing. Convolutional neural network (CNN) is the most popular neural network model to handle image analysis and can recognize patterns from an image accurately. This study detected Drusens, Optic Disc Cupping, and Tessellation diseases using 534 fundus images. The architecture used Convolutional neural network-based Support Vector Machine (CNN based SVM) and DenseNet, which is a Convolutional neural network architecture development. In obtaining the best results, in this study, we use several variations of the optimizer, namely adam, nadam, and RMSprop, and the best results from this study can be seen from the accuracy value of 93,21% using the DenseNet architecture.

关键词： Support vector machines Visualization Biomedical optical imaging optical computing Retina optical imaging Data models

来源：评论

学校读者我要写书评

暂无评论

Photonic pattern reconstruction enabled by on-chip online learning and inference

引用

JOURNAL OF PHYSICS-PHOTONICS 2021年第2期3卷

作者： Marquez, Bicky A. Guo, Zhimu Morison, Hugh Shekhar, Sudip Chrostowski, Lukas Prucnal, Paul Shastri, Bhavin J. Queens Univ Dept Phys Engn Phys & Astron Kingston ON KL7 3N6 Canada Univ British Columbia Dept Elect & Comp Engn Vancouver BC V6T 1Z4 Canada Princeton Univ Dept Elect Engn Princeton NJ 08544 USA

Recent investigations in neuromorphic photonics exploit optical device physics for neuron models, and optical interconnects for distributed, parallel, and analog processing. Integrated solutions enabled by silicon photonics enable high-bandwidth, low-latency and low switching energy, making it a promising candidate for special-purpose artificial intelligence hardware accelerators. Here, we experimentally demonstrate a silicon photonic chip that can perform training and testing of a Hopfield network, i.e. recurrent neural network, via vector dot products. We demonstrate that after online training, our trained Hopfield network can successfully reconstruct corrupted input patterns.

关键词： neuromorphic photonics brain-inspired computing photonic integrated circuits recurrent neural network artificial intelligence hardware

来源：评论

学校读者我要写书评

暂无评论

一种自适应的分布式深度神经网络推理框架

引用

建模与仿真 2024年第4期13卷 4449-4459页

作者：吴宾宾杨桂松上海理工大学光电信息与计算机工程学院上海

近年来,随着深度学习的发展,深度神经网络(Deep neural network,DNN)模型变得越来越复杂,所需的内存和数据传输量也随之增大,这不仅降低了DNN的训练和推理速度,也限制了DNN在一些内存较小、计算能力较差的物联网(Internet of Things,IoT... 详细信息

近年来,随着深度学习的发展,深度神经网络(Deep neural network,DNN)模型变得越来越复杂,所需的内存和数据传输量也随之增大,这不仅降低了DNN的训练和推理速度,也限制了DNN在一些内存较小、计算能力较差的物联网(Internet of Things,IoT)设备上的部署。现有研究将基于云–边–端协同的分布式计算框架与深度神经网络相结合,组成了分布式深度神经网络(Distributed Deep neural network,DDNN)框架,该框架在IoT应用场景下有着显著的优势。然而,DDNN框架存在设备的计算能力有限、以及设备之间的传输成本较高等问题。针对上述问题,本文提出了自适应的分布式深度神经网络(Adaptive Distributed Deep neural network,ADA-DDNN)推理框架。ADA-DDNN框架采用了多个边缘出口,这些边缘出口允许ADA-DDNN框架中的模型在不同的深度层次上进行自适应地推理,以适应不同的任务需求和数据特性。此外,该框架增加了额外的边缘处理模块,边缘处理模块可以在边缘端进行特征融合之前,判断每个终端模块的输出结果是否可信,若可信,则直接输出分类结果,无需进行特征融合和后续计算。这大大增加了样本的边缘出口概率,减少了后续的计算成本。本文在开放的CIFAR-10数据集上进行验证,实验结果表明,ADD-DDNN框架在保证云端测试精度的前提下,显著提升了边缘测试精度。

关键词：深度神经网络物联网分布式深度神经网络特征融合

来源：评论

学校读者我要写书评

暂无评论

Sonar Image Generation from Underwater Optic Images utilizing Sensor models and Depth Estimation with neural network

Sonar Image Generation from Underwater Optic Images utilizin...

引用

OCEANS

作者： Minsung Sung Son-Cheol Yu Pohang University of Science and Technology (POSTECH) Pohang Republic of Korea Division of Advanced Nuclear Engineering Pohang University of Science and Technology (POSTECH) Pohang Republic of Korea

ISBN: (纸本)9781665427883

This paper proposes a method to generate a sonar image from an optical image, called optic-to-sonar which is one direction of opti-acoustic translation. To convert an optical image into a sonar image having a different geometry, the proposed method first reconstructs a three-dimensional point cloud by estimating the depth information using a neural network from given optical images. Then the sonar image is generated through the sonar projection model. We verified the proposed method through water tank experiments. The proposed method can generate a sonar image of the same viewpoint as a given optical image. So, it can be utilized as a dataset augmentation technique to further develop opti-acoustics in an underwater environment where data acquisition is challenging.

关键词： Image synthesis neural networks Data acquisition Sonar Estimation optical computing optical fiber networks

来源：评论

学校读者我要写书评

暂无评论

neural Distortion Fields for Spatial Calibration of Wide Field-of-View Near-Eye Displays

arXiv

引用

arXiv 2022年

作者： Hiroi, Yuichi Someya, Kiyosato Itoh, Yuta The University of Tokyo 7-3-1 Hongo Bunkyo-ku Tokyo113-8654 Japan Tokyo Institute of Technology 4259 Nagatsutacho Midoriku Kanagawa Yokohama226-0026 Japan

We propose a spatial calibration method for wide Field-of-View (FoV) Near-Eye Displays (NEDs) with complex image distortions. Image distortions in NEDs can destroy the reality of the virtual object and cause sickness. To achieve distortion-free images in NEDs, it is necessary to establish a pixel-by-pixel correspondence between the viewpoint and the displayed image. Designing compact and wide-FoV NEDs requires complex optical designs. In such designs, the displayed images are subject to gaze-contingent, non-linear geometric distortions, which explicit geometric models can be difficult to represent or computationally intensive to optimize. To solve these problems, we propose neural Distortion Field (NDF), a fully-connected deep neural network that implicitly represents display surfaces complexly distorted in spaces. NDF takes spatial position and gaze direction as input and outputs the display pixel coordinate and its intensity as perceived in the input gaze direction. We synthesize the distortion map from a novel viewpoint by querying points on the ray from the viewpoint and computing a weighted sum to project output display coordinates into an image. Experiments showed that NDF calibrates an augmented reality NED with 90◦ FoV with about 3.23 pixel (5.8 arcmin) median error using only 8 training viewpoints. Additionally, we confirmed that NDF calibrates more accurately than the non-linear polynomial fitting, especially around the center of the FoV. Copyright © 2022, The Authors. All rights reserved.

关键词： Pixels

来源：评论

学校读者我要写书评

暂无评论

Improved all photonics diffraction neural network based on multi-channel integrated optical fibers

引用

ISCIENCE 2025年第6期28卷 112596页

作者： Zhu, Jiakang An, Qichang Yang, Fei Chinese Acad Sci Changchun Inst Opt Fine Mech & Phys Changchun 130033 Peoples R China Univ Chinese Acad Sci Beijing 100039 Peoples R China

AI's exponentially growing computational demands conflict with slow hardware advances. The high-power consumption and long training times of large-scale models call for alternative solutions. optical computing-based traditional optical networks and diffractive deep neural network (D2NN) still face deployment challenges and reliance on electronic networks. To address these issues, we replace the free-space interlayer propagation in conventional optical networks with fiber-based propagation. This preserves the advantages of traditional optical networks while providing additional benefits such as ease of deployment, reduced dependence on electronic networks, and enhanced robustness. Experimental results demonstrate that this untrained structure exhibits strong nonlinear mapping capabilities across different configurations, yielding distinct outputs for three input targets, especially at 1550 nm. Furthermore, the influence of environment and noise is around 1% in target recognition. Leveraging inherent spectral discrimination, this architecture enables multidimensional target identification with important implications for complex target classification and multidimensional sensing.

关键词： Physics Optics Applied computing

来源：评论

学校读者我要写书评

暂无评论

Privacy-preserving Social Distance Monitoring on Microcontrollers with Low-Resolution Infrared Sensors and CNNs

arXiv

引用

arXiv 2022年

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

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 9.38-68.57µJ. © 2022, CC BY-NC-ND.

关键词： Edge computing

来源：评论

学校读者我要写书评

暂无评论

Enhancements of Model and Method in Lithography Hotspot Identification

Enhancements of Model and Method in Lithography Hotspot Iden...

引用

Design, Automation and Test in Europe Conference and Exhibition (DATE)

作者： Huang, Xuanyu Zhang, Rui Huang, Yu Wang, Peiyao Li, Mei Tsinghua Univ Ctr Nano & Micro Mech Dept Mech Engn Beijing Peoples R China HiSilicon Technol Co Ltd Shenzhen Peoples R China

ISBN: (纸本)9783981926354

The manufacturing of integrated circuits (ICs) has been continuously improved through the advancement of fabrication technology nodes. However the lithography hotspots (HSs) caused by optical diffraction problems seriously affect the yield and reliability of ICs. Although lithography simulation can accurately capture the HSs through physically simulating the lithography process, it requires a lot of computing resources, which usually takes > 100 CPU .h/mm(2) [1]. Due to the image recognition nature, the state-of-the-art HS identification algorithms based on deep learning have obvious advantages in reducing run time comparing to the traditional algorithms. However, its accuracy still needs to be enhanced since there are many false alarms of non-hotspots (NHSs) and escapes of the real HSs, which makes it difficult to be a signoff technique. In this paper, we propose two enhancements in HS identification. First, a hybrid deep learning model is proposed in lithography HS identification, which includes a CNN model to combine physical features. Second, an ensemble learning method is proposed based on multiple sub-models. The proposed enhanced model and method can achieve high HS identification accuracy on the benchmarks 1-4 of the ICCAD 2012 dataset with recall> 98.8%. In addition, it can achieve even 100% recall on the benchmark 1 and benchmark 3 while maintaining the precision at a high level with 53.6% and 87.1%, respectively. Moreover, for the first time it can achieve not only 100% recall on benchmark 5, but also high precision of 61.8%, which is much higher than any published deep learning methods for HSs identification, as far as we know. The proposed model and methodology can be applied in industrial IC designs due to its effectiveness and efficiency.

关键词： lithography hotspot ensemble learning deep learning convolutional neural network

来源：评论

学校读者我要写书评

暂无评论

Investigating Self-Attention in Swin-Unet Model for Disc and Cup Segmentation

Investigating Self-Attention in Swin-Unet Model for Disc and...

引用

International Conference on Information Technology, Computer and Electrical Engineering (ICITACEE)

作者： Jehua Kusuma Dewa Ema Rachmawati Gamma Kosala School of Computing Telkom University Bandung Indonesia

Glaucoma is a condition of the eyes that results from damage to the optic nerve, which could potentially lead to loss of vision if not addressed promptly. Because glaucoma rarely shows early symptoms in the sufferer, it requires regular observation of the retinal fundus by an ophthalmologist to find out if this eye disorder appears. Doctors’ observations are subjective, so they are inconsistent and take a long time. As a result, a computer-aided diagnostic (CAD) system was created that can automate the examination of retinal fundus images to detect glaucoma in its initial phase. In a consistent and time-saving manner by using optic disc and cup segmentation and cup-to-disc ratio (CDR) computation. CAD systems can also be used as decision support by doctors. Several previous image segmentation studies have proposed using convolution neural network (CNN) and Vision Transformer (ViT) based models and their combinations. However, the encoder-decoder model based on CNN is large and is slow in computation. The ViT model has the problem that the computational amount of the model increases when the image size also increases. Therefore, the segmentation method uses a Swin Transformer-based encoder-decoder model, Swin-Unet, which has the advantage of a self-attention mechanism performed on local windows and has linear computation. This paper presents a case study of optical disc and cup segmentation using the Swin-Unet method with the REFUGE dataset. The vCDR calculation with a threshold of 0.63 yields an accuracy of 94%. The IoU score results using the REFUGE dataset resulted in a score of 84% for the disc part and 80% for the cup part.

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：