检索结果-内蒙古大学图书馆

Towards property-preserving JPEG encryption with structured permutation and adaptive group differentiation

Visual computer 2024年第9期40卷 6421-6447页

作者： Su, Guo-Dong Chang, Ching-Chun Lin, Chia-Chen Chang, Chin-Chen School of Big Data and Artificial Intelligence Fujian Polytechnic Normal University Fuzhou350300 China Department of Information Engineering and Computer Science Feng Chia University Taichung40724 Taiwan Fujian Province University Fuzhou China Department of Computer Science and Information Engineering National Chin-Yi University of Technology Taichung41170 Taiwan

The security and privacy of digital images are a major concern in cyberspace. JPEG is the most widely used image compression standard and yet there are problems with format compatibility and file size preservation in most of the modern encryption schemes for JPEG images. To address these problems, we propose a novel properties-preserving encryption scheme that obfuscates JPEG images by scrambling AC and DC coefficients. The key observation is that an innovative structured permutation has a remarkable effect on enciphering the coefficients and simultaneously prevents value overflow as well as size increment. On this basis, we first leverage the row-wise reordering and reversing technique to disorganize the DC coefficients. Afterward, a series of groups of consecutive DC differential values are produced using adaptive group differentiation and each of them is independently permuted inside. Finally, we perform block-wise ACCs permutation and intra-block ACCs shuffling to further confuse the outline of the original image. Moreover, invariant features regarding the coefficients are incorporated in the key generation process to resist various cryptanalytic attacks. Experimental results demonstrate a significant improvement in terms of format compatibility and file size preservation, and verify the security against various attacks. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2023.

关键词： Cryptography

来源：评论

学校读者我要写书评

暂无评论

Scalable Algorithms for Densest Subgraph Discovery 39

Scalable Algorithms for Densest Subgraph Discovery

引用

39th IEEE International Conference on data Engineering, ICDE 2023

作者： Luo, Wensheng Tang, Zhuo Fang, Yixiang Ma, Chenhao Zhou, Xu Hunan University College of Computer Science and Electronic Engineering China The Chinese University of HongKong School of Data Science Shenzhen China

ISBN: (纸本)9798350322279

As a fundamental problem in graph data mining, Densest Subgraph Discovery (DSD) aims to find the subgraph with the highest density from a graph. It has been studied for several decades and found a large number of real-world applications, such as network community detection, regulatory motif discovery in DNA, graph index construction, and fake follower detection. Although there are many existing DSD algorithms, they are often not scalable or efficient to process large-scale graphs, since most of them are serial algorithms and can only leverage the computing resource of a single CPU core. To tackle these issues, in this paper we propose efficient parallel algorithms for solving the DSD problems on both undirected and directed graphs at scale. Our main idea is to use the k-cores (a kind of dense subgraph) to approximate the densest subgraph in the undirected graphs, and then propose efficient parallel algorithms for computing the cores by optimizing the iterative process and also reducing the number of iterations. We further extend this idea for directed graphs by introducing a novel concept, named w-induced subgraph, to avoid unnecessary enumerations of x or y when searching [x,y]-cores (a kind of directed dense subgraph to approximate the densest). To verify the scalability and efficiency of the proposed algorithms, we have conducted extensive experiments on 12 large real-world graphs, and four of them are billion-scale. The experimental results show that our proposed algorithms outperform the state-of-the-art algorithms on both undirected and directed graphs, in terms of scalability and efficiency. © 2023 IEEE.

关键词： Directed graphs

来源：评论

学校读者我要写书评

暂无评论

SYSTEM THINKING MEETS data science/ENGINEERING 67

SYSTEM THINKING MEETS DATA SCIENCE/ENGINEERING

引用

67th Annual Meeting of the International Society for the Systems sciences, ISSS 2023

作者： Malele, Vusumuzi Unit for Data Science and Computing School of Computer Science and Information Systems North-West University Vanderbijlpark Campus South Africa

ISBN: (纸本)9781713893455

With the advent of data 3.0 and analytics 3.0, system thinkers are in the position to provide a bigger picture in data science and data engineering. In the data life cycle, a system thinking approach emphasises data-driven decision-making. A System Thinker approaches problem-solving by viewing the problems as part of a wider, data-resourced and dynamic system, and a data Practitioner supports the data life cycle by collecting, transforming, and analyzing data, and communicating results to inform and guide decision-making. This paper uses explanatory research and a pragmatic case study approach to look at the (i) What is the role of system thinking and data science/engineering skill in data-driven decision-making or organisation? (ii) Does the combination of system thinking and data science/engineering give rise to a new discipline? (iii) What are the skills needed in this new discipline? The research shows that the system thinking skills in the data life cycle are important. System thinkers meet data practitioners to provide a bigger picture of data-driven decision-making. The latter ascertains the position of a system thinker in any industrial revolution (i.e. industry 4.0. and industry 5.0). Furthermore, a data-System Thinker is proposed as a new career field. © Copyright, Author (Creative Commons), 2023.

关键词： Systems thinking

来源：评论

学校读者我要写书评

暂无评论

Modelling High-Order Social Relations for Item Recommendation (Extended Abstract) 39

Modelling High-Order Social Relations for Item Recommendatio...

引用

39th IEEE International Conference on data Engineering, ICDE 2023

作者： Liu, Yang Chen, Liang He, Xiangnan Peng, Jiaying Zheng, Zibin Tang, Jie Sun Yat-sen University School of Computer Science and Engineering Guangzhou China University of Science and Technology of China School of Data Science School of Information Science and Technology Hefei China Tsinghua University Department of Computer Science and Technology Beijing China

ISBN: (纸本)9798350322279

Personalized recommendation is becoming increasingly important in online information systems in the current era of information explosion. In real-world scenarios, when a user considers which items to consume, the decision choice may be affected by her friends. For example, she may ask her friends for suggestions or be attracted by products purchased by one friend. As such, to provide satisfactory recommendation service, it is important to account for the evidence in social relations when they are available to use. Several prior efforts have been made to leverage social relations to build the recommender system and verified their utility. However, most existing methods, such as the well-known TrustSVD, leverage only first-order social relations, i.e., the direct neighbors that are connected to the target user. The high-order social relations, e.g., the friends of friends, which are very informative to reveal user preference, have been largely ignored. © 2023 IEEE.

关键词： Online systems

来源：评论

学校读者我要写书评

暂无评论

Extraction and Classification of Medicines from Handwritten Medical Prescriptions 18

Extraction and Classification of Medicines from Handwritten ...

引用

18th IEEE International Conference on Emerging Technologies, ICET 2023

作者： Zia, Javeria Habib, Usman Naeem, Muhammad Asif National University of Computer and Emerging Sciences Department of Ai & Data Science Islamabad Pakistan

ISBN: (纸本)9798350328172

In the modern world of rapid technological progress and digital transformation, industries are swiftly shifting from paper-based to digital systems. Document digitization, especially for image-based documents lacking machine readability, is gaining global attention. Sectors like real estate, finance, law, healthcare, and education are actively adopting digital strategies. In healthcare, where handwritten prescriptions lead to inefficiencies such as errors in medicine name by pharmacist due to doctor's hand writing lead to catastrophic failure and delays in various processes. Therefore, digital solutions are being developed to enhance patient care. This study proposes an approach to extract and categorize medicines from handwritten prescriptions, improving accuracy and communication with pharmacists. We proposed an approach that uses stroke enhancement apart from pre-processing with comprehensive multi-modal architecture, including a transformer based deep neural network and a sequence-to-sequence model such as DONUT and layoutLMv2 model. Our approach achieves a mean accuracy of about 0.85%. It surpasses models like LayoutLMv2 and DONUT by providing a balance in accuracy and consistency. Furthermore, the extracted features are passed to the various state of art machine learning classification algorithms upon which Gradient Boosting, Random forest and Ridge classifier produced the promising results for medicine type classification. © 2023 IEEE.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

An Amalgamated CNN-Transformer Network for Lightweight Image Super-Resolution

Journal of Network Intelligence

引用

Journal of Network Intelligence 2024年第3期9卷 1376-1387页

作者： Fang, Jinsheng Lin, Hanjiang Zeng, Kun School of Computer Science and Engineering Minnan Normal University Zhangzhou363000 China Key Laboratory of Data Science and Intelligence Application Fujian Province University Fujian Zhangzhou363000 China School of Computer and Control engineering Fujian Provincial Key Laboratory of Information Processing and Intelligent Control Minjiang University Fuzhou350108 China

Recently, Transformer-based methods for single image super-resolution (SISR) have achieved better performance advantages than the methods based on convolutional neural network (CNN). Exploiting self-attention mechanism to model global context definitely improves the SR results. However, the neglect of local information will bring inevitable reduction of the network performance. In this work, we propose an Amalgamated CNN-Transformer network for lightweight SR, namely ACTSR. Specifically, an amalgamated CNN-Transformer block (ACTB) is developed to extract the useful information of both local and global features. By employing stacked ACTBs, our ACTSR extracts more informative deep features beneficially for super-resolution reconstruction to improve network performance while keeps lightweight and flexible enough. Extensive experiments on commonly used benchmark datasets validate our ACTSR outperforms the advanced competitors. Our codes are available at: https://***/ginsengf/ACTSR. © 2024, Taiwan Ubiquitous Information CO LTD. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

A Trichotomy for Transductive Online Learning 37

A Trichotomy for Transductive Online Learning

引用

37th Conference on Neural Information Processing Systems, NeurIPS 2023

作者： Hanneke, Steve Moran, Shay Shafer, Jonathan Department of Computer Science Purdue University United States Faculty of Mathematics Faculty of Computer Science Faculty of Data and Decision Sciences Technion - Israel Institute of Technology Israel Computer Science Division UC Berkeley United States

ISBN: (纸本)9781713899921

We present new upper and lower bounds on the number of learner mistakes in the 'transductive' online learning setting of Ben-David, Kushilevitz and Mansour (1997). This setting is similar to standard online learning, except that the adversary fixes a sequence of instances x1, ..., xn to be labeled at the start of the game, and this sequence is known to the learner. Qualitatively, we prove a trichotomy, stating that the minimal number of mistakes made by the learner as n grows can take only one of precisely three possible values: n, Θ (log(n)), or Θ(1). Furthermore, this behavior is determined by a combination of the VC dimension and the Littlestone dimension. Quantitatively, we show a variety of bounds relating the number of mistakes to well-known combinatorial dimensions. In particular, we improve the known lower bound on the constant in the Θ(1) case from Ω (Equation presented)(plog(d) ) to Ω(log(d)) where d is the Littlestone dimension. Finally, we extend our results to cover multiclass classification and the agnostic setting. © 2023 Neural information processing systems foundation. All rights reserved.

关键词： E-learning

来源：评论

学校读者我要写书评

暂无评论

Prediction of Alzheimer’s Disease Using Adaptive Fine-Tuned Deep Resnet-50 with Attention Mechanism

引用

SN computer science 2024年第4期5卷 392页

作者： Venkatesh, R. Anantharajan, Shenbagarajan Gunasekaran, Shenbagalakshmi Yogaraja, C.A. Gethzi Ahila Poornima, I. Computer Science and Engineering Ramco Institute of Technology Rajapalayam Tamil Nadu Virudhunagar 626117 India Artificial Intelligence and Data Science Mepco Schlenk Engineering College Sivakasi Tamilnadu Virudhunagar 626005 India Computer Science and Engineering Mepco Schlenk Engineering College Sivakasi Tamilnadu Virudhunagar 626005 India

Alzheimer’s disease (AD) sufferers can benefit from early detection and diagnosis so that protective treatments can be taken before permanent brain impairment develops. Magnetic resonance imaging (MRI)-detected lesion sites in AD patients exhibit significant heterogeneity and these characteristics are spread throughout the imaging space, according to cross-sectional imaging studies of Alzheimer's disease. Because conventional Deep Learning (DL) methods can’t generate enough long-distance correlations in the characteristic space, adaptive fine-tuned deep resnet-50 with attention mechanism (AFDR-AM) is introduced in this research for the early identification of AD. The input MRI data is de-noised using a wavelet thresholding approach, and the contrast is enhanced using Improved Color Histogram Equalization (ICHE). Image recognition and classification characteristics are extracted at the feature extraction phase using a Gabor Filter Bank (GFB). To narrow down a big list of traits, the Artificial Bee Colony Optimization (ABCO) technique can be used to pick the most useful ones and get rid of the rest. Measures of accuracy, precision, recall, f1-score, specificity, and MSE are used to evaluate the effectiveness of the suggested strategy. Comprehensive experiments reveal that our approach achieved the best performance compared to other approaches. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd 2024.

关键词： AFDR-AM Alzheimer’s disease (AD) Artificial Bee colony optimization (ABCO) Deep learning (DL) Gabor filter bank (GFB) Improved color histogram equalization (ICHE) Magnetic resonance imaging (MRI)

来源：评论

学校读者我要写书评

暂无评论

Deep Masked Input UNet Framework based Lips Segmentation prediction for Gender Classification 2

Deep Masked Input UNet Framework based Lips Segmentation pre...

引用

2nd IEEE International Conference on Advances in Information Technology, ICAIT 2024

作者： Devi, M. Shyamala Hepzhibha Rachel, S. Janani, G. Irene, B. Joy Praisy, E. Humaira, S. Department of Computer Science and Engineering Panimalar Engineering College Tamilnadu Chennai India Department of Artificial Intelligence and Data Science Panimalar Engineering College Tamilnadu Chennai India

ISBN: (纸本)9798350383867

Assessments have demonstrated that the human lip and its motions provide a wealth of knowledge about the identity and substance of communication. However, due to large differences in illumination condition, head perspective, and background, obtaining strong and precise lip image segmentation in natural settings remains difficult. This paper recommends Deep Masked Input UNet that categorizes the Gender based on the lip segmentation with high precision. For implementation, 10,132 face images from the KAGGLE Lip Segmentation dataset were used. The dataset includes 5066 images of the face and 5066 segmented lip images. The proposed Deep Masked Input UNet starts by masking the original picture with a segmented lip image to create masked face images with lips segments. Utilizing the Relu Activation function, Deep Masked Input UNet with contracting and expansion was enabled. The masked lip images are fitted to the proposed Deep Masked Input UNet and traditional CNN models. The results show that the suggested Deep Masked Input UNet model performs better in lip segmentation and gender classification, with a high accuracy of 98.95%. © 2024 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Derandomizing Multi-Distribution Learning 38

Derandomizing Multi-Distribution Learning

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Larsen, Kasper Green Montasser, Omar Zhivotovskiy, Nikita Department of Computer Science Aarhus University Denmark Department of Statistics and Data Science Yale University United States Department of Statistics University of California Berkeley United States

Multi-distribution or collaborative learning involves learning a single predictor that works well across multiple data distributions, using samples from each during training. Recent research on multi-distribution learning, focusing on binary loss and finite VC dimension classes, has shown near-optimal sample complexity that is achieved with oracle efficient algorithms. That is, these algorithms are computationally efficient given an efficient ERM for the class. Unlike in classical PAC learning, where the optimal sample complexity is achieved with deterministic predictors, current multi-distribution learning algorithms output randomized predictors. This raises the question: can these algorithms be derandomized to produce a deterministic predictor for multiple distributions? Through a reduction to discrepancy minimization, we show that derandomizing multi-distribution learning is computationally hard, even when ERM is computationally efficient. On the positive side, we identify a structural condition enabling an efficient black-box reduction, converting existing randomized multi-distribution predictors into deterministic ones. © 2024 Neural information processing systems foundation. All rights reserved.

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：