检索结果-内蒙古大学图书馆

An image retrieval system using an artificial neural network for the identification and classification of lung diseases

引用

AIP conference Proceedings 2025年第1期3254卷

作者： Atul Pratap Singh Ajeet Singh Amit Kumar Hari Om Sharan C. S. Raghuvanshi Department of AI Noida Institute of Engineering and Technology Greater Noida UP India Dept of CSE Moradabad Institute of Technology Moradabad UP India Dept of CSE FET Rama University Kanpur UP India

Automated systems for lung disease identification and classification are being developed with quick pace, thanks to the rapid advancements in the field of Medical Imaging Technologies. This paper describes the latest development and application of a novel Artificial Neural Network (ANN)-based image retrieval system to improve the precision and efficacy in terms of lung disease diagnosis. The approach under consideration is designed to tackle the difficulties involved in accurately identifying and categorizing lung disorders using medical imaging, including CT and X-rays. By utilizing Artificial Neural Networks (ANNs), more especially convolutional neural networks (CNNs), the system seeks to identify complex patterns and features in the images that are frequently invisible to the human observers. This makes it possible for the system to pick up on discriminative representations of both healthy lung structure and different disease states. There are several steps involved in the system’s design. First, a large collection of annotated lung photos is assembled, including a wide variety of lung conditions and the associated healthy states. A pre-processing pipeline is then put into place to standardize the images, guaranteeing a constant quality and making feature extraction easier. Then, in order to enable efficient learning and classification, the CNN architecture is painstakingly built, paying close attention to layer configurations, activation functions, and optimization techniques. Additionally, the system uses image retrieval algorithms, which allow for the effective search and retrieval of pertinent medical images from the database in response to query inputs. In order to promote appropriate diagnosis and treatment planning, this retrieval functionality helps medical practitioners locate similar cases for comparative study and reference. Comprehensive tests are carried out utilizing benchmark datasets to assess the system’s performance, and metrics including accura

关键词：

来源：评论

学校读者我要写书评

暂无评论

基于多注意融合网络的输煤皮带异物识别方法

引用

西安科技大学学报 2024年第5期44卷 976-984页

作者：李利梁晶陈旭东寇发荣潘红光西安科技大学电气与控制工程学院陕西西安710054 西安市电气设备状态监测与供电安全重点实验室陕西西安710054 西安科技大学机械工程学院陕西西安710054

为改善现有输煤皮带异物识别算法网络参数量大、识别精度不高的问题,及时避免大块煤和矸石、锚杆等带来的安全隐患,提出了一种基于多注意融合网络的输煤皮带异物识别方法,使用低照度图像处理算法对数据集进行预处理,采用融合局部注意力... 详细信息

为改善现有输煤皮带异物识别算法网络参数量大、识别精度不高的问题,及时避免大块煤和矸石、锚杆等带来的安全隐患,提出了一种基于多注意融合网络的输煤皮带异物识别方法,使用低照度图像处理算法对数据集进行预处理,采用融合局部注意力残差块作为基本特征提取单元,在残差块中融入带有额外偏移量的可变形卷积以增加对不规则特征的描述,用注意力机制对全局特征图做期望最大化处理。结果表明:在Cifar 10数据集和矿用皮带传输异物识别数据集的识别准确率分别为93.7%和84.8%;与ShufflenetV2、MobileNetV2、ResNet 50、ResNet 110、Darknet 53算法相比,识别准确率分别提升了4.7%、3.9%、0.4%、0.5%、1.7%;与识别准确率相近的ResNet 50、ResNet 110算法相比,网络参数量和计算复杂度大大减小。识别方法能够快速识别输煤皮带异物,且具有较高的识别准确率,对保障煤矿运输系统的安全运行具有参考意义。

关键词：异物识别输煤皮带 Darknet网络可变形卷积注意力机制

来源：评论

学校读者我要写书评

暂无评论

Modified-PBIL Based User Selection for Multi-user Massive MIMO systems with Massive Connectivity

Modified-PBIL Based User Selection for Multi-user Massive MI...

引用

IEEE conference on Computer Communications Workshops, INFOCOM Wksps

作者： Jing Jiang Junyu Chen Yongbin xie Hongjiang Lei Ling Zheng Shaanxi Key Laboratory of Information Communication Network and Security Xi'an University of Posts and Telecommunication Xi'an China Chongqing University of Posts and Telecommunications Chongqing China

ISBN: (数字)9781728186955

ISBN: (纸本)9781728186962

Inspired by the big data processing capability of the machine learning, we propose a user selection algorithm based on modified population-based incremental learning (MPBIL) for multi-user massive MIMO systems with massive connectivity. With the objective of enhancing the algorithm efficiency, the proposed algorithm evolves the population exploiting both the superior individuals and the best individual. In further, we design the orthogonal permutation to increase individual diversity and avoid the overfit of the classical population-based incremental learning. Simulation results demonstrate that the performance of the proposed algorithm is far better than the classical greedy based user selection method while maintaining low complexity, especially for the large number of MU-MIMO users and candidates.

关键词： Sociology Statistics MIMO communication Convergence Machine learning algorithms Computational complexity

来源：评论

学校读者我要写书评

暂无评论

Research on Face Detection Technology Based on MTCNN

Research on Face Detection Technology Based on MTCNN

引用

International conference on Computer Network, Electronic and Automation (ICCNEA)

作者： Ning Zhang Junmin Luo Wuqi Gao School of Computer Science and Engineering Xi'an Technological University Xi'an China

ISBN: (数字)9781728170831

ISBN: (纸本)9781728170848

Face detection is an important research direction in the field of target detection. For the input image, the position of the face is returned. In order to complete the task of face detection using deep learning, data input, feature extraction and face feature detection are three steps, among which feature extraction is the most important part. By studying the basic principles of current mainstream target detection algorithms, this paper compares the characteristics of Two-stage and One-stage detection models and their application in face detection tasks. At the same time, MTCNN(Multi-task convolution neural network) is deeply analyzed and its implementation principle is introduced in detail. The real effect of MTCNN in face detection task is verified by experiments. The results of the model are compared with those of yolov3 model in the wider face dataset.

关键词： Faces Feature extraction Object detection Training Face detection Task analysis Operating systems

来源：评论

学校读者我要写书评

暂无评论

Machine Learning in Information Visualization: Using Rule-based Learning algorithms to Personalize Interfaces

Machine Learning in Information Visualization: Using Rule-ba...

引用

International conference on systems, Signals and image processing, IWSSIP

作者： Leonardo Souza Silva Ricardo Nakamura Fatima L. S. Nunes Interactive Technologies Laboratory Escola Politécnica Universidade de São Paulo São Paulo Brasil Laboratory of Computer Applications for Health Care Escola de Artes Ciências e Humanidades Universidade de São Paulo São Paulo Brasil

ISBN: (数字)9781728175393

ISBN: (纸本)9781728175409

Machine Learning has offered great potential in areas such as data mining and information retrieval, while less attention has been given to its use in the design of Information Visualization tools. Based on user needs, a properly designed interface can enhance augmented perception; however, the project is usually conducted without the inclusion of users. Machine Learning techniques can be useful to promote the users' involvement in the interface development process, contributing to offer personalized visualizations. Based on a dataset from a User Experience experiment with an interactive 3D environment to visualize medical temporal data, we present an approach using rule-based learning algorithms to promote the user's involvement. Our results indicate that Association Rule Learning can be a valuable resource to improve the design of Information Visualization interface tools as well as to customize visualization according to the user's preferences.

关键词： Data visualization image color analysis Tools Visualization Itemsets Three-dimensional displays Shape

来源：评论

学校读者我要写书评

暂无评论

GPU-based digital image correlation system for real-time strain-controlled fatigue and strain field measurement 11

GPU-based digital image correlation system for real-time str...

引用

conference on Optical Measurement systems for Industrial Inspection xi held at SPIE Optical Metrology Symposium

作者： Blug, A. Regina, D. J. Eckmann, S. Senn, M. Eberl, C. Bertz, A. Carl, D. Fraunhofer Inst Phys Measurement Tech IPM Freiburg Germany Fraunhofer Inst Mech Mat IWM Freiburg Germany

ISBN: (数字)9781510627925

ISBN: (纸本)9781510627925

This article reports a novel GPU-based 2D digital image correlation system (2D-DIC) overcoming two major limitations of this technique: It measures marker-free, i.e. without sample preparation, and the sampling rate meets the recommendations of ASTM E606. The GPU implementation enables zero-normalized cross correlation (ZNCC) calculation rates of up to 25 kHz for 256 x 256 pixel ROIs. This high-speed image processing system is combined with a high-resolution telecentric lens observing a 10 mm field-of-view, coaxial LED illumination, and a camera acquiring 2040 x 256 pixel images with 1.2 kHz. The optics resolve the microstructure of the surface even of polished cylindrical steel specimen. The displacement uncertainty is below 0.5 mu m and the reproducibility in zero-strain tests approximately 10(-5) (1 sigma) of the field-of-view. For strain-controlled testing, a minimum of two displacement subsets per image are evaluated for average strain with a sampling rate of 1.2 kHz. Similar to mechanical extensometers, an analogue 0-10V displacement signal serves as a feedback for standard PID controllers. The average latency is below 2 ms allowing for cycle frequencies up to 10 Hz. For strain-field measurement, the number of ROIs limits the frame rate, e.g., the correlation rate of 25 kHz is sufficient to evaluate 10 images per second with 2500 ROIs each. This frame rate is still sufficient to compare the maximum and minimum strain fields within a cycle in real-time, e.g. for crack detection. The result is a marker-free and non-contact DIC sensor suitable for both strain-controlled fatigue testing and real-time full-field strain evaluation.

关键词： digital image correlation materials testing experimental mechanics real-time image processing closed-loop control graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

Optimized Implementation of Modified Gram Schmidt Algorithm on VLIW Architecture

Optimized Implementation of Modified Gram Schmidt Algorithm ...

引用

IEEE 4th World conference on Complex systems (WCCS)

作者： Najoui, Mohamed Hatim, Anas Belkouch, Said Chabini, Noureddine Mohamed V Univ ENSET Rabat E2SN Lab Rabat Morocco Ibn Zohr Univ ENSA Agadir LGEMS Lab Agadir Morocco Cadi Ayyad Univ ENSA Marrakech LGECOS Lab Marrakech Morocco Royal Mil Coll Canada Dept Elect & Comp Engn Kingston ON Canada

ISBN: (纸本)9781728112329

Modified Gram Schmidt (MGS) is one of the well-known forms of QR decomposition (QRD) algorithms. It has been used in many signal and image processing applications to solve least square problem, linear equations or to invert matrices. Nevertheless, QRD is considered a computationally expensive operation, and its sequential implementation doesn't meet the requirements of many real time applications. In this paper, we propose an optimized MGS algorithm version based on software pipelining and loop unrolling techniques. The suggested MGS version is parallel and well suited for VLIW architectures. The implementation is done under TI C6678 VLIW DSP and the obtained results show great improvements against the standard MGS and the optimized vendor QRD implementations.

关键词： Signal and image processing MGS VLIW Software Pipelining DSP

来源：评论

学校读者我要写书评

暂无评论

Comprehensive study of Deep learning based Telugu OCR

Comprehensive study of Deep learning based Telugu OCR

引用

International conference on Smart systems and Inventive Technology (ICSSIT)

作者： Muni Sekhar Velpuru Priyadarshini Chatterjee G Tejasree M Ravi Kumar S Nageswara Rao Vardhaman College of Engineering Hyderabad India JBREC Hyderabad India

ISBN: (数字)9781728158211

ISBN: (纸本)9781728158228

Optical Character Recognition (OCR) is a technology to recognize text in the images. images in the sense means scanned documents or pictures. With the help of an OCR, it is possible to convert any image text into machine-readable text data. OCR has varied uses in understanding any scripts written in any language. They also convert the text with high accuracy. After any image text has gone through OCR processing, they edited with the help of MS Office. With the advent of OCR, there is no need to digitize the image text. Another advanced technology that is extensively used in OCR is deep earning. Deep learning can be called as a subfield of Machine Learning. This area is inspired by the structure and function of the brain. Deep learning has many algorithms that are used to solve problems. This area offers scalable solutions to problems. Another very important benefit of deep learning is feature extraction from unorganized data. In a broader sense, this area provides automated feature learning. It is called “deep”, as the layers of the network are made to learn deep. As segmentation is a part of OCR, delve into it in this paper. Segmentation is a subpart of image processing; OCR and it goes hand in hand with deep learning. Segmentation helps in easier analysis of the parts of an image. Nowadays segmentation is clubbed with deep learning and artificial intelligence to analyze and understand those features that was not possible attest a decade ago. Deep learning helps to learn the input patterns so that they can predict object classes. image segmentation divides the input image into segments to help us to do a better analysis of the image.

关键词： Optical character recognition software image segmentation Machine learning Character recognition Optical imaging Ciphers Classification algorithms

来源：评论

学校读者我要写书评

暂无评论

Remote Sensing for Maritime Monitoring and Vessel Prompt Identification 11th

Remote Sensing for Maritime Monitoring and Vessel Prompt Ide...

引用

11th International conference on Multimedia and Network Information systems (MISSI)

作者： Reggiannini, Marco Righi, Marco Tampucci, Marco Bedini, Luigi Di Paola, Claudio Martinelli, Massimo Mercurio, Costanzo Salerno, Emanuele Natl Res Council Italy Inst Informat Sci & Technol I-56124 Pisa Italy Mapsat SRL Contrada Piano Cappelle 1 I-82100 Benevento BN Italy

ISBN: (纸本)9783319986784;9783319986777

The main purpose of the work described in this paper concerns the development of a platform dedicated to sea surveillance, capable of detecting and identifying illegal maritime traffic. This platform results from the cascade implementation of several image processing algorithms that take as input Radar or Optical maps captured by satelliteborne sensors. More in detail, the processing chain is dedicated to (i) the detection of vessel targets in the input map, (ii) the refined estimation of the vessel most descriptive geometrical features and, finally, (iii) the estimation of the kinematic status of the vessel. This platform will represent a new tool for combating unauthorized fishing, irregular migration and related smuggling activities.

关键词： Maritime traffic monitoring SAR sensing Optical sensing Ship detection image segmentation image classification Wake detection and analysis

来源：评论

学校读者我要写书评

暂无评论

Application of Auto-Threshold plugin for cut surface evaluation of white cheese in brine 8

Application of Auto-Threshold plugin for cut surface evaluat...

引用

8th International Scientific conference on Engineering, Technologies and systems, TechSys 2019

作者： Bosakova-Ardenska, Atanaska Panayotov, Peter Boyanova, Petya Department of Computer Systems and Technologies University of Food Technologies Plovdiv Bulgaria Department of Milk and Dairy Products Technology University of Food Technologies Plovdiv Bulgaria

This paper presents a research over application of global thresholding methods for evaluation of cut surface of white cheese in brine. Six trademarks Bulgarian white cheese in brine produced of cow, sheep and goat milk are used for cut surface evaluation. Examined cheese samples are evaluated by experts and the results are used as referent assessment for results of image processing trough thresholding algorithms. A Java program is developed for image segmentation with specific thresholding algorithm using two or three colors in different modes to produce segmented image. The images of cheese samples (their cut surfaces) are processed with developed program using all of the seventeen thresholding algorithms which are implemented in Auto-Threshold plugin for open source software imageJ in different modes. A correlation analysis shows that some of thresholding algorithms which are implemented in Auto-Threshold plugin could be used for cut surface evaluation of white cheese in brine. © 2019 IOP Publishing Ltd.

关键词： image segmentation

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：