检索结果-内蒙古大学图书馆

Hyperspectral image restoration using noise gradient and dual priors under mixed noise conditions

CAAI Transactions on Intelligence Technology 2025年第1期10卷 72-93页

作者： Hazique Aetesam Suman Kumar Maji V.B.Surya Prasath Computer Science and Engineering Birla Institute of Technology MesraBiharIndia Computer Science and Engineering Indian Institute of Technology PatnaBiharIndia Department of Computer Science University of CincinnatiCincinnatiOhioUSA

Images obtained from hyperspectral sensors provide information about the target area that extends beyond the visible portions of the electromagnetic ***,due to sensor limitations and imperfections during the image acquisition and transmission phases,noise is introduced into the acquired image,which can have a negative impact on downstream analyses such as classification,target tracking,and spectral *** in hyperspectral images(HSI)is modelled as a combination from several sources,including Gaussian/impulse noise,stripes,and *** HSI restoration method for such a mixed noise model is ***,a joint optimisation framework is proposed for recovering hyperspectral data corrupted by mixed Gaussian-impulse noise by estimating both the clean data as well as the sparse/impulse noise ***,a hyper-Laplacian prior is used along both the spatial and spectral dimensions to express sparsity in clean image ***,to model the sparse nature of impulse noise,anℓ_(1)−norm over the impulse noise gradient is *** the proposed methodology employs two distinct priors,the authors refer to it as the hyperspectral dual prior(HySpDualP)*** the best of authors'knowledge,this joint optimisation framework is the first attempt in this *** handle the non-smooth and nonconvex nature of the generalℓ_(p)−norm-based regularisation term,a generalised shrinkage/thresholding(GST)solver is ***,an efficient split-Bregman approach is used to solve the resulting optimisation *** results on synthetic data and real HSI datacube obtained from hyperspectral sensors demonstrate that the authors’proposed model outperforms state-of-the-art methods,both visually and in terms of various image quality assessment metrics.

关键词： hyper-laplacian prior hyperspectral images image restoration mixed noise variational approach

来源：评论

学校读者我要写书评

暂无评论

Enhancing Predictive Capabilities for Cyber Physical Systems Through Supervised Learning

Informatica (Slovenia)

引用

Informatica (Slovenia) 2025年第16期49卷 77-86页

作者： Dhanalakshmi, B. Tamije Selvy, P. Department of Computer Science and Engineering Dr.N.G. P Institute of technology India Department of Computer Science and Engineering Hindusthan College of Engineering and Technology India

The rapid advancement and proliferation of Cyber-Physical Systems (CPS) have led to an exponential increase in the volume of data generated continuously. Efficient classification of this streaming data is crucial for predicting system behaviors and enabling proactive decision-making. This research aims to extract actionable knowledge from the continuous data streams of CPS and predict their behavior using advanced supervised learning algorithms. The predictions facilitate timely interventions and necessary actions within the interconnected physical network. The background of this work lies in the intersection of CPS, machine learning, and data stream mining. Traditional batch processing methods are inadequate for real-time analysis of CPS data due to their inherent latency and computational inefficiency. This research employs state-of-the-art techniques for real-time data processing, including incremental learning, sliding window models, and ensemble methods tailored for streaming data. Our approach differs from existing works by focusing on a comprehensive framework that integrates real-time data ingestion, preprocessing, feature extraction, and model updating in a seamless pipeline. Unlike previous studies that often rely on static datasets and offline analysis, our method ensures continuous learning and adaptation to evolving data patterns. Comparative analysis with existing techniques demonstrates superior performance in terms of accuracy, latency, and scalability. Specifically, our models achieved an average classification accuracy of 92%, with a precision of 90%, recall of 89%, and an F1 score of 89.5%. These metrics indicate significant improvements over traditional batch processing methods, which typically lag in responsiveness and adaptability. This research provides a robust and efficient solution for the real-time classification of streaming data from CPS, enhancing the system's ability to predict behaviors and take necessary actions promptly. © 2025 Slov

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Residual-enhanced graph convolutional networks with hypersphere mapping for anomaly detection in attributed networks

Data Science and Management

引用

Data science and Management 2025年第2期8卷 137-146页

作者： Khan, Wasim Mohd, Afsaruddin Suaib, Mohammad Ishrat, Mohammad Shaikh, Anwar Ahamed Faisal, Syed Mohd Department of Computer Science and Engineering Koneru Lakshmaiah Education Foundation Vaddeswaram 522502 India Department of Computer Engineering & Applications GLA University Mathura281406 India Department of Computer Science and Engineering Integral University Lucknow226026 India

In the burgeoning field of anomaly detection within attributed networks, traditional methodologies often encounter the intricacies of network complexity, particularly in capturing nonlinearity and sparsity. This study introduces an innovative approach that synergizes the strengths of graph convolutional networks with advanced deep residual learning and a unique residual-based attention mechanism, thereby creating a more nuanced and efficient method for anomaly detection in complex networks. The heart of our model lies in the integration of graph convolutional networks that capture complex structural relationships within the network data. This is further bolstered by deep residual learning, which is employed to model intricate nonlinear connections directly from input data. A pivotal innovation in our approach is the incorporation of a residual-based attention mechanism. This mechanism dynamically adjusts the importance of nodes based on their residual information, thereby significantly enhancing the sensitivity of the model to subtle anomalies. Furthermore, we introduce a novel hypersphere mapping technique in the latent space to distinctly separate normal and anomalous data. This mapping is the key to our model's ability to pinpoint anomalies with greater precision. An extensive experimental setup was used to validate the efficacy of the proposed model. Using attributed social network datasets, we demonstrate that our model not only competes with but also surpasses existing state-of-the-art methods in anomaly detection. The results show the exceptional capability of our model to handle the multifaceted nature of real-world networks. © 2024 Xi'an Jiaotong University

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Agricultural supply chain management using hyperledger and AIOT

引用

Journal of Ambient Intelligence and Humanized Computing 2025年第4期16卷 471-485页

作者： Jha, Anurag Kumar Raj, Aparna Jha, Ashish Kumar Shetty, Sujala D. Department of Computer Science Bits-Pilani Dubai Campus Dubai United Arab Emirates Department of Electrical and Computer Engineering Carnegie Mellon University PittsburghPA United States

Supply chain management and Hyperledger are two interconnected domains. They leverage blockchain technology to enhance efficiency, transparency, and security in supply chain operations. Together, they provide a decentralized, traceable, and real-time platform for recording and managing transactions. This combination is particularly valuable for industries dealing with sensitive goods, as it provides accurate traceability and real-time information. This paper explores the integration of supply chain management with Hyperledger blockchain technology to enhance efficiency, transparency, and security in supply chain operations. We propose a decentralized Hyperledger Fabric blockchain network to improve traceability, security, and efficiency by monitoring environmental conditions. This approach is particularly beneficial for transporting sensitive goods, such as medical supplies and perishable items, by ensuring optimal conditions and real-time data accessibility. The integration of Artificial Intelligence (AI) further enhances insights, reduces waste, and improves overall efficiency. By utilizing a distributed network free from third-party intermediaries, the system ensures immutability and remote accessibility, addressing challenges related to transporting heat and humidity sensitive products. Our experimental assessment demonstrates the benefits of private blockchain technologies, including enhanced security, regulatory compliance, compatibility, flexibility, and scalability. This study presents a detailed methodology for developing a traceable, efficient, and sustainable agricultural supply chain. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： Efficiency

来源：评论

学校读者我要写书评

暂无评论

Bengali reduplication generation with finite-state transducers (FSTs)

引用

International Journal of Speech Technology 2024年第3期27卷 729-737页

作者： Barman, Abhijit Saha, Diganta Pal, Alok Ranjan Department of Computer Science and Engineering Jadavpur University Kolkata700032 India Department of Computer Science and Engineering College of Engineering and Management Kolaghat721171 India

Reduplication is a highly productive process in Bengali word formation, with significant implications for various natural language processing (NLP) applications, such as parts-of-speech tagging and sentiment analysis. Despite its importance, this area has not been extensively explored in computational linguistics, especially for low-resource languages like Bengali. This study first demonstrates that a two-way finite-state transducer (FST) can effectively capture complete reduplication generation processes in Bengali. Second, it is shown that the formation of partial reduplication requires a set of 2-way FSTs due to the diverse patterns involved in Bengali partial reduplications. Third, the research highlights the utility of the reduplication generation process in identifying Bengali reduplication instances, achieving a commendable F1-Score of 88.11%. This method outperforms current state-of-the-art methods for identifying reduplicated expressions in Bengali text. This research contributes valuable insights into the computational representation of reduplication in Bengali, offering potential enhancements for NLP tasks in low-resource language scenarios. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Privacy Comparison for Bitcoin Light Client Implementations 6

Privacy Comparison for Bitcoin Light Client Implementations

引用

6th Conference on Advances in Financial Technologies, AFT 2024

作者： Kotzer, Arad Rottenstreich, Ori The Department of Computer Science Technion Haifa Israel The Department of Computer Science The Department of Electrical and Computer Engineering Technion Haifa Israel

ISBN: (纸本)9783959773454

Light clients implement a simple solution for Bitcoin’s scalability problem, as they do not store the entire blockchain but only the state of particular addresses of interest. To be able to keep track of the updated state of their addresses, light clients rely on full nodes to provide them with the required information. To do so, they must reveal information about the addresses they are interested in. This paper studies the two most common light client implementations, SPV and Neutrino with regards to their privacy. We define privacy metrics for comparing the privacy of the different implementations. We evaluate and compare the privacy of the implementations over time on real Bitcoin data and discuss the inherent privacy-communication tradeoff. In addition, we propose general techniques to enhance light client privacy in the existing implementations. Finally, we propose a new SPV-based light client model, the aggregation model, evaluate it, and show it can achieve enhanced privacy than in the existing light client implementations. © Arad Kotzer and Ori Rottenstreich.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

MMInstruct: a high-quality multi-modal instruction tuning dataset with extensive diversity

引用

science China(Information sciences) 2024年第12期67卷 36-51页

作者： Yangzhou LIU Yue CAO Zhangwei GAO Weiyun WANG Zhe CHEN Wenhai WANG Hao TIAN Lewei LU Xizhou ZHU Tong LU Yu QIAO Jifeng DAI School of Computer Science Nanjing University School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University Shanghai AI Laboratory School of Computer Science Fudan University Department of Information Engineering The Chinese University of Hong Kong SenseTime Research Department of Electronic Engineering Tsinghua University

Despite the effectiveness of vision-language supervised fine-tuning in enhancing the performance of vision large language models(VLLMs), existing visual instruction tuning datasets include the following limitations.(1) Instruction annotation quality: despite existing VLLMs exhibiting strong performance,instructions generated by those advanced VLLMs may still suffer from inaccuracies, such as hallucinations.(2) Instructions and image diversity: the limited range of instruction types and the lack of diversity in image data may impact the model's ability to generate diversified and closer to real-world scenarios outputs. To address these challenges, we construct a high-quality, diverse visual instruction tuning dataset MMInstruct,which consists of 973k instructions from 24 domains. There are four instruction types: judgment, multiplechoice, long visual question answering, and short visual question answering. To construct MMInstruct, we propose an instruction generation data engine that leverages GPT-4V, GPT-3.5, and manual correction. Our instruction generation engine enables semi-automatic, low-cost, and multi-domain instruction generation at 1/6 the cost of manual construction. Through extensive experiment validation and ablation experiments,we demonstrate that MMInstruct could significantly improve the performance of VLLMs, e.g., the model fine-tuning on MMInstruct achieves new state-of-the-art performance on 10 out of 12 benchmarks. The code and data shall be available at https://***/yuecao0119/MMInstruct.

关键词： instruction tuning multi-modal multi-domain dataset vision large language model

来源：评论

学校读者我要写书评

暂无评论

A novel interpolation consistency for bad generative adversarial networks (IC-BGAN)

引用

Multimedia Tools and Applications 2024年第38期83卷 86161-86205页

作者： Iraji, Mohammad Saber Tanha, Jafar Balafar, Mohammad-Ali Feizi-Derakhshi, Mohammad-Reza Department of Computer Engineering Faculty of Electrical and Computer Engineering University of Tabriz Tabriz Iran

Semi-supervised learning techniques utilize both labeled and unlabeled images to enhance classification performance in scenarios where labeled images are limited. However, challenges such as integrating unlabeled images with incorrect pseudo-labels, determining appropriate thresholds for the pseudo-labels, and label prediction fluctuations on low-confidence unlabeled images, hinder the effectiveness of existing methods. This research introduces a novel framework named Interpolation Consistency for Bad Generative Adversarial Networks (IC-BGAN) that utilizes a new loss function. The proposed model combines bad adversarial training, fusion techniques, and regularization to address the limitations of semi-supervised learning. IC-BGAN creates three types of image augmentations and label consistency regularization in interpolation of bad fake images, real and bad fake images, and unlabeled images. It demonstrates linear interpolation behavior, reducing fluctuations in predictions, improving stability, and facilitating the identification of decision boundaries in low-density areas. The regularization techniques boost the discriminative capability of the classifier and discriminator, and send a better signal to the bad generator. This improves the generalization and the generation of diverse inter-class fake images as support vectors with information near the true decision boundary, which helps to correct the pseudo-labeling of unlabeled images. The proposed approach achieves notable improvements in error rate from 2.87 to 1.47 on the Modified National Institute of Standards and Technology (MNIST) dataset, 3.59 to 3.13 on the Street View House Numbers (SVHN) dataset, and 12.13 to 9.59 on the Canadian Institute for Advanced Research, 10 classes (CIFAR-10) dataset using 1000 labeled training images. Additionally, it reduces the error rate from 22.11 to 18.40 on the CINIC-10 dataset when using 700 labeled images per class. The experiments demonstrate the IC-BGAN framework outp

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Enhancing thyroid disease prediction with improved XGBoost model and bias management techniques

引用

Multimedia Tools and Applications 2025年第16期84卷 16757-16788页

作者： Dalal, Surjeet Lilhore, Umesh Kumar Faujdar, Neetu Simaiya, Sarita Agrawal, Akshat Rani, Uma Mohan, Anand Department of Computer Science and Engineering Amity University Haryana Gurugram India Department of Computer Science and Engineering Galgotias University Uttar Pradesh Greater Noida India Department of Computer Engineering and Applications GLA University Uttar Pradesh Mathura India Department of Computer Science & Engineering Amity University Haryana Gurugram India Department of Computer Science and Engineering World College of Technology & Management Haryana Gurugram India Department of Physics Kunwar Singh College Bihar Darbhanga India

The thyroid gland, a pivotal regulator of essential physiological functions, orchestrates the production and release of thyroid hormones, playing a vital role in metabolism, growth, development, and overall bodily functions. While thyroid nodules are predominantly benign, certain cases may manifest malignancy, leading to thyroid cancer. Despite the generally favorable prognosis, early detection and intervention remain paramount for optimal outcomes. Given the limitations in clinical data, this paper introduces an innovative approach employing a modified XGBoost model that integrates both clinical and molecular properties into a unified predictive model. Utilizing fine-needle aspiration biopsies and gene expression analysis, our model aims to enhance cancer prediction in thyroid nodules. We explore the influence of incorporating high-quality clinical data on the prioritized gene set, employing a nature-inspired method for hyperparameter optimization. The modified XGBoost model emerges as highly recommended for non-parametric feature selection, extraction, and prediction due to its remarkable accuracy. Our proposed model attains an impressive 94.6% accuracy, surpassing other machine learning models such as C5.0 (92.5%), CART (88.7%), CHAID (83.4%), Quest (81.5%), LSVM (77%), and random tree (24.1%). The real-time efficiency of our model is demonstrated in the results section. Experiments showcase the superior performance of combined molecular and clinical data-based prediction models. Emphasis is placed on the importance of scrutinizing the origins of clinical data and maintaining vigilance over variable quality. This study signifies a substantial advancement in thyroid cancer prediction, offering a robust and accurate model that integrates diverse data sources, ultimately contributing to improved patient outcomes. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

MLRT-UNet:An Efficient Multi-Level Relation Transformer Based U-Net for Thyroid Nodule Segmentation

引用

computer Modeling in engineering & sciences 2025年第4期143卷 413-448页

作者： Kaku Haribabu Prasath R Praveen Joe IR Department of Computer Science and Engineering RMK College of Engineering and TechnologyTiruvallur601206India School of Computer Science and Engineering Vellore Institute of TechnologyChennai600127India

Thyroid nodules,a common disorder in the endocrine system,require accurate segmentation in ultrasound images for effective diagnosis and ***,achieving precise segmentation remains a challenge due to various factors,including scattering noise,low contrast,and limited resolution in ultrasound *** existing segmentation models have made progress,they still suffer from several limitations,such as high error rates,low generalizability,overfitting,limited feature learning capability,*** address these challenges,this paper proposes a Multi-level Relation Transformer-based U-Net(MLRT-UNet)to improve thyroid nodule *** MLRTUNet leverages a novel Relation Transformer,which processes images at multiple scales,overcoming the limitations of traditional encoding *** transformer integrates both local and global features effectively through selfattention and cross-attention units,capturing intricate relationships within the *** approach also introduces a Co-operative Transformer Fusion(CTF)module to combine multi-scale features from different encoding layers,enhancing the model’s ability to capture complex patterns in the ***,the Relation Transformer block enhances long-distance dependencies during the decoding process,improving segmentation *** results showthat the MLRT-UNet achieves high segmentation accuracy,reaching 98.2% on the Digital Database Thyroid Image(DDT)dataset,97.8% on the Thyroid Nodule 3493(TG3K)dataset,and 98.2% on the Thyroid Nodule3K(TN3K)*** findings demonstrate that the proposed method significantly enhances the accuracy of thyroid nodule segmentation,addressing the limitations of existing models.

关键词： Thyroid nodules endocrine system multi-level relation transformer U-Net self-attention external attention co-operative transformer fusion thyroid nodules 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：