检索结果-内蒙古大学图书馆

The 3rd International Conference on Distributed Sensing and Intelligent systems (ICDSIS 2022)

作者： M. Tarek M. Meribout v. K. Tiwari Department of Electrical and Computer Engineering Khalifa University Abu Dhabi UAE

With the rapid progress of tomography systems, a need has emerged for innovation that can provide high accuracy measurements, fast acquisition rate, and low-cost robust systems for multi-phase flow and medical instruments. The main objective of this work is to design and implement a 3D Electrical Impedance Tomography (EIT) system to effectively monitor a multi-phase flow which consists of an annular flow comprising of an air-core of unknown diameter, surrounded by a liquid phase, such as oil and water mixture of high water-cut value. The system consists of an array of electrode sensors evenly distributed around the pipe confining the three-phase flow. Subsequently, a dedicated hardware module (GPU) is used to reconstruct these signals, captured from the sensors, into a 3D image of the flow by solving the forward-inverse problem. The annular flow is generated using a flow conditioner consisting of a swirl cage followed by a bluff body that generates vortices that push the high dense liquid phase to the outer side of the pipeline. The 3D EIT system is designed to accurately measure the gas liquid fraction and the liquid-liquid fraction and provide the 3D volume image of the flow. Hence, it can be utilized for pipeline and reservoirs characterization to achieve high efficiency, cost reduction, and safer production. The approach is to design a dedicated, highly parallel hardware architecture that can sustain the high computation complexity of the 3D EIT algorithm. Coincidentally, advanced software development tools are available to help effectively develop these parallel hardware algorithms. The main novelty is utilizing GPUs to outperform the existing real-time tomography hardware platforms, mainly based on Field Programmable Gate Array (FPGAs), which have the disadvantage of being less dense and feature lower precise arithmetic operations. Matrix multiplication and matrix inversion are considered the most intensive computation in Electric Impedance Tomography (EIT). T

关键词： computational complexity computerised instrumentation electric impedance imaging flow measurement flowmeters image reconstruction inverse problems matrix multiplication multi-threading parallel processing

来源：评论

学校读者我要写书评

暂无评论

LMM-vQA: Advancing video Quality Assessment with Large Multimodal Models

引用

IEEE Transactions on Circuits and systems for video Technology 2025年

作者： Ge, Qihang Sun, Wei Zhang, Yu Li, Yunhao Ji, Zhongpeng Sun, Fengyu Jui, Shangling Min, Xiongkuo Zhai, Guangtao Shanghai Jiao Tong University Institute of Image Communication and Information Processing Shanghai 200240 China East China Normal University School of Communication and Electronic Engineering Shanghai 200241 China Shanghai Jiao Tong University MoE Key Laboratory of Artificial Intelligence AI Institute Shanghai 200240 China Huawei Technologies Shanghai 200040 China

The explosive growth of videos on streaming media platforms has underscored the urgent need for effective video quality assessment (vQA) algorithms to monitor and perceptually optimize the quality of streaming videos. However, vQA remains an extremely challenging task due to the diverse video content and the complex spatial and temporal distortions, thus necessitating more advanced methods to address these issues. Nowadays, large multimodal models (LMMs), such as GPT-4v, have exhibited strong capabilities for various visual understanding tasks, motivating us to leverage the powerful multimodal representation ability of LMMs to solve the vQA task. Therefore, we propose an Large Multi-Modal based video Quality Assessment (LMM-vQA) model, which introduces a novel spatiotemporal visual modeling strategy for quality-aware feature extraction. Specifically, we reformulate the quality regression problem into a question and answering (Q&A) task and construct Q&A prompts for vQA instruction tuning. Then, we design a spatiotemporal vision encoder to extract spatial and temporal features to represent the quality characteristics of videos, which are subsequently mapped into the language space by the spatiotemporal projector for modality alignment. Finally, the aligned visual tokens and the quality-inquired text tokens are aggregated as inputs for the large language model (LLM) to generate the quality score as well as the quality level. Extensive experiments demonstrate that LMM-vQA achieves state-of-the-art performance across five vQA benchmarks, exhibiting an average improvement of 5% in generalization ability over existing methods. Furthermore, due to the advanced design of the spatiotemporal encoder and projector, LMM-vQA also performs exceptionally well on general video understanding tasks, further validating its effectiveness. Our code will be released at https://***/Sueqk/LMM-vQA. © 1991-2012 IEEE.

关键词： Damage detection

来源：评论

学校读者我要写书评

暂无评论

Feature-reduction fuzzy co-clustering approach for hyper-spectral image analysis

引用

KNOWLEDGE-BASED systems 2021年 216卷 106549-106549页

作者： Nha van Pham Long The Pham Pedrycz, Witold Long Thanh Ngo Le Quy Don Tech Univ Dept Informat Syst Hanoi Vietnam MIST Inst Sci & Technol 17 Hoang Sam Hanoi Vietnam Univ Alberta Dept Elect & Comp Engn Edmonton AB T6R 2V4 Canada

Fuzzy co-clustering algorithms are the effective techniques for multi-dimensional clustering in which all features are considered of equal importance (relevance). In fact, the features' importance could be different, even several of them could be considered redundant. The removal of the redundant features has formed the idea of feature-reduction in problems of the big data processing. In this paper, we propose a new unsupervised learning scheme by incorporating the feature-weighted entropy into the objective function of fuzzy co-clustering, called the Feature-Reduction Fuzzy Co-Clustering Algorithm (FRFCoC). First, a new objective function is formed on the basis of the original fuzzy co-clustering objective function which adds parameters representing the entropy weight of the different features. Next, a feature-reduction and clustering automatic schema are adjusted based on FCoC's original learning schema which calculates new parameters and conditions to eliminate irrelevant feature components. FRFCoC algorithm can be mathematically shown to converge after a finite number of iterations. The experiment results were conducted on some many-features data sets and hyperspectral images that have demonstrated the outstanding performance of FRFCoC algorithm compared with some previously proposed algorithms. (C) 2020 Elsevier B.v. All rights reserved.

关键词： Fuzzy co-clustering Dimensionality reduction Cluster tendency Hyper-spectral satellite image Land-cover classification

来源：评论

学校读者我要写书评

暂无评论

TabNAS: rejection sampling for neural architecture search on tabular datasets 22

TabNAS: rejection sampling for neural architecture search on...

引用

Proceedings of the 36th International Conference on Neural Information processing systems

作者： Chengrun Yang Gabriel Bender Hanxiao Liu Pieter-Jan Kindermans Madeleine Udell Yifeng Lu Quoc v. Le Da Huang Google Research Brain Team Stanford University

ISBN: (纸本)9781713871088

The best neural architecture for a given machine learning problem depends on many factors: not only the complexity and structure of the dataset, but also on resource constraints including latency, compute, energy consumption, etc. Neural architecture search (NAS) for tabular datasets is an important but under-explored problem. Previous NAS algorithms designed for image search spaces incorporate resource constraints directly into the reinforcement learning (RL) rewards. However, for NAS on tabular datasets, this protocol often discovers suboptimal architectures. This paper develops TabNAS, a new and more effective approach to handle resource constraints in tabular NAS using an RL controller motivated by the idea of rejection sampling. TabNAS immediately discards any architecture that violates the resource constraints without training or learning from that architecture. TabNAS uses a Monte-Carlo-based correction to the RL policy gradient update to account for this extra filtering step. Results on several tabular datasets demonstrate the superiority of TabNAS over previous reward-shaping methods: it finds better models that obey the constraints.

关键词：

来源：评论

学校读者我要写书评

暂无评论

The identification of doubly stochastic circular image model 24

The identification of doubly stochastic circular image model

引用

24th KES International Conference on Knowledge-Based and Intelligent Information and Engineering systems, KES 2020

作者： Krasheninnikov, victor Malenova, Olga Subbotin, Alexey Ulyanovsk State Technical University 32 Severny Venets Ulyanovsk432027 Russia

In some practical situations, images are set on a circle. For example, images of the facies (thin film) of dried biological fluid, eyes, cut of a tree trunk, etc. Currently, most of the image processing works deal with images defined on rectangular two-dimensional grids or grids of higher dimension. The features of circle images require their consideration in their mathematical models. In this paper, an autoregressive models of homogeneous and inhomogeneous random fields defined on a circle are considered as representations of images with radial or radial-circular structure. In the present paper, autoregressive models of circular images are considered. To represent heterogeneous images with random heterogeneities, «doubly stochastic» models are used in which one or more images control the parameters of the resulting image. Pseudo-gradient algorithms for the modal identification are proposed. The conducted statistical modeling showed that these algorithms give good model identification. © 2020 The Authors. Published by Elsevier B.v.

关键词： Stochastic models

来源：评论

学校读者我要写书评

暂无评论

Optimization of stereoscopic imager performance by computer simulation of geometrical calibration using optical design software

引用

OPTICS EXPRESS 2019年第13期27卷 17819-17839页

作者： Gorevoy, Alexey, v Machikhin, Alexander S. Batshev, vladislav, I Kolyuchkin, vasiliy Ya Russian Acad Sci Sci & Technol Ctr Unique Instrumentat Lab Acoustoopt Spect 15 Butlerova Str Moscow 117342 Russia Natl Res Univ Moscow Power Engn Univ Inst Automat & Comp Engn Dept Elect Engn & Introscopy 14 Krasnokazarmermaya Str Moscow 111250 Russia Bauman Moscow State Tech Univ Dept Radioelect & Laser Technol 5 2 Ya Baumanskaya Str Moscow 105005 Russia

Stereoscopic imagers are widely used in machine vision for three-dimensional (3D) visualization as well as in non-destructive testing for quantitative characterization of cracks, delamination and other defects. Measurement capability in these systems is provided by a proper combination of the optical parameters and data processing techniques. Conventional approach to their design consists of two sequential stages: optical system design and optimization of calibration and image processing algorithms. Such two-stage procedure often complicates both the hardware and the software, and results in a time-ineffective design procedure and cost-ineffective solution. We demonstrate a more effective approach and propose to estimate errors of 3D measurements at the early optical design stage. We show that computer simulation using optical design software allows not only optimizing optical parameters of the imager but also choosing the most effective mathematical model of the system and the equipment necessary for calibration. We tested the proposed approach on the design of miniature prism-based stereoscopic system and analyzed the impact of various factors (aberrations, tolerances, etc.) as on the image quality, so on the quality of calibration and 3D measurements accuracy. The proposed joint design approach may be highly effective for various measurement systems and applications when both optical parameters and image processing algorithms are not defined in advance and are necessary to be optimized. (C) 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

关键词： Camera calibration Digital image processing image processing algorithms image quality Optical design software Optical systems

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Applied for Histological Diagnosis of Breast Cancer

引用

IEEE ACCESS 2020年 8卷 162432-162448页

作者： Yari, Yasin Nguyen, Thuy v. Nguyen, Hieu T. Univ South Eastern Norway Dept Sci & Ind Syst N-3616 Kongsberg Norway Posts & Telecommun Inst Technol Fac Informat Technol Hanoi 122 Vietnam

Deep learning, as one of the currently most popular computer science research trends, improves neural networks, which has more and deeper layers allowing higher abstraction levels and more accurate data analysis. Although deep convolutional neural networks, as a deep learning algorithm, has recently achieved promising results in data analysis, the requirement for a large amount of data prevents its use in medical data analysis since it is challenging to obtain data from the medical field. Breast cancer is a common cancer in women. To diagnose this kind of cancer, breast cell shapes in histopathology images should be examined by senior pathologists. The number of pathologists per population in the world is not enough, especially in Africa, and human mistake may occur in diagnosis procedure. After the evaluation of deep learning methods and algorithms in breast histological data processing, we tried to improve the current systems' accuracy. As a result, this study proposes two effective deep transfer learning-based models, which rely on pre-trained DCNN using a large collection of imageNet dataset images that improve current state-of-the-art systems in both binary and multiclass classification. We transfer pre-trained weights of the ResNet50 and DesneNet121 on the imagenet as initial weights and fine-tune these models with a deep classifier with data augmentation to detect various malignant and benign samples tissues in the two categories of binary classification and multiclass classification. The proposed models have been examined with optimized hyperparameters in magnification-dependent and magnification-independent classification modes. In the multiclass classification, the proposed system achieved up to 98% accuracy. As for binary classification, the proposed system provides up to 100% accuracy. The results outperform previous studies accuracies in all defined performance metrics in breast cancer CAD systems from histological images.

关键词： Feature extraction Breast cancer Solid modeling Machine learning Biomedical imaging Breakhis dataset breast cancer CNN computer-aided diagnosis (CAD) medical image classification densNet ResNet

来源：评论

学校读者我要写书评

暂无评论

Algorithm for stabilizing video images

Algorithm for stabilizing video images

引用

2020 Wave Electronics and its Application in Information and Telecommunication systems, WECONF 2020

作者： Sotnikov, A.v. Shipatov, A.v. Myo, Aung Kyaw Fedorov, P.A. Fedorov, A.R. National Research University of Electronic Technology Institute of Systems and Software Engineering and Information Technology Moscow Russia

The basic methods for stabilizing video images are considered. The requirements for the image stabilization algorithm in the video streaming processing system for unmanned aerial vehicles (UAvs) are formulated. An improved modification of the video stabilization algorithm is proposed based on a comparison of key image points. As a descriptor of the found local features, the modern high-performance FREAK method was applied, surpassing existing algorithms both in the speed of generation and comparison of feature vectors and in the accuracy of comparison and resistance to image transformations. This made it possible to significantly increase the performance of the stabilization algorithm and improve the quality of stabilized video images due to a more accurate calculation of the parameters of inter-frame transformations, which, in turn, depends on the accuracy of matching sets of singular points. The results of testing the proposed algorithm on real and synthesized video sequences are presented. © 2020 IEEE.

关键词： video streaming

来源：评论

学校读者我要写书评

暂无评论

Cosmic Ray rejection with attention augmented deep learning

引用

ASTRONOMY AND COMPUTING 2022年第0期40卷

作者： Bhavanam, S. R. Channappayya, S. S. Srijith, P. K. Desai, S. IIT Hyderabad Dept Elect Engn Kandi 502285 Telangana India IIT Hyderabad Dept Comp Sci & Engn Kandi 502285 Telangana India IIT Hyderabad Dept Phys Kandi 502285 Telangana India

Cosmic Ray (CR) hits are the major contaminants in astronomical imaging and spectroscopic observa-tions involving solid-state detectors. Correctly identifying and masking them is a crucial part of the image processing pipeline, since it may otherwise lead to spurious detections. For this purpose, we have developed and tested a novel Deep Learning based framework for the automatic detection of CR hits from astronomical imaging data from two different imagers: Dark Energy Camera (DECam) and Las Cumbres Observatory Global Telescope (LCOGT). We considered two baseline models namely deepCR and Cosmic-CoNN, which are the current state-of-the-art learning based algorithms that were trained using Hubble Space Telescope (HST) ACS/WFC and LCOGT Network images respectively. We have experimented with the idea of augmenting the baseline models using Attention Gates (AGs) to improve the CR detection performance. We have trained our models on DECam data and demonstrate a consistent marginal improvement by adding AGs in True Positive Rate (TPR) at 0.01% False Positive Rate (FPR) and Precision at 95% TPR over the aforementioned baseline models for the DECam dataset. We demonstrate that the proposed AG augmented models provide significant gain in TPR at 0.01% FPR when tested on previously unseen LCO test data having images from three distinct telescope classes. Furthermore, we demonstrate that the proposed baseline models with and without attention augmentation outperform state-of-the-art models such as Astro-SCRAPPY, Maximask (that is trained natively on DECam data) and pre-trained ground-based Cosmic-CoNN. This study demonstrates that the AG module augmentation enables us to get a better deepCR and Cosmic-CoNN models and to improve their generalization capability on unseen data. (C) 2022 Elsevier B.v. All rights reserved.

关键词： image processing CCD observation Cosmic rays Neural networks Attention gates

来源：评论

学校读者我要写书评

暂无评论

复合主干融合的图像增强算法

引用

计算机系统应用 2024年第4期33卷 194-201页

作者：万嘉龙况立群曹亚明郭磊熊风光中北大学计算机科学与技术学院太原030051 机器视觉与虚拟现实山西省重点实验室太原030051 山西省视觉信息处理及智能机器人工程研究中心太原030051

基于深度学习的微光图像增强算法所生成的图像普遍存在噪声凸显和细节丢失等问题,而端对端深度学习算法的性能又在很大程度上依赖于骨干网络的提取能力,因此,通过探索更有效的骨干网络结构可以提升微光增强任务的性能收益.本文提出了一... 详细信息

基于深度学习的微光图像增强算法所生成的图像普遍存在噪声凸显和细节丢失等问题,而端对端深度学习算法的性能又在很大程度上依赖于骨干网络的提取能力,因此,通过探索更有效的骨干网络结构可以提升微光增强任务的性能收益.本文提出了一种复合主干网络融合策略的图像增强算法,将不同图像增强算法中的主干网络进行融合,以提高整体网络的特征提取能力.该算法通过逐层融合来自不同主干网络的特征信息,将复合特征引导到解码器中,再充分利用不同的上采样方法,将主干网络融合的特征进行堆叠,最终生成正常光照条件下的图像.通过与现有的主流算法进行定量与定性的对比实验,结果显示,本文方法显著提升了微光图像的亮度,同时保留图像的细节特征,在峰值信噪比和结构相似性客观指标上,在LOL-v2数据集上达到了24.35 dB和0.871,有效解决了图像增强后的噪声凸显和细节丢失问题.

关键词：网络融合图像增强微光复合主干

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：