检索结果-内蒙古大学图书馆

1st ACM Workshop on Security Implications of Deepfakes and Cheapfakes, WDC 2022, co-located with ACM AsiaCCS 2022

作者： Park, Geon-Woo Park, Eun-Ju Woo, Simon S. Computer Science & Engineering Dept. Sunkyunkwan University Suwon Korea Republic of Applied Data Science Dept. Sunkyunkwan University Suwon Korea Republic of

ISBN: (纸本)9781450391788

With the rapid growth of deep learning methods, AI technologies for generating deepfake videos also have been significantly advanced. Nowadays, the manipulated videos such as deepfakes are so sophisticated that one cannot easily differentiate between real and fake, and one can create such videos with little effort. However, such technologies can be likely to be abused by people with malicious intents. To address this issue, approaches and efforts to detect deepfakes have been researched significantly. However, the performances of the detectors in general depends on the quantity and quality of the training data. In this paper, we introduce a new deepfake dataset, Zoom-DF, which can be injected during the remote meeting and video conferencing, to create a sequence of fake participant images. While most deepfake datasets focus on the face area, our dataset primarily targets for the remote meeting, and manipulates movements of the participants. We evaluate existing deepfake detectors on our new Zoom-DF dataset and present the performance results. © 2022 ACM.

关键词： Video conferencing

来源：评论

学校读者我要写书评

暂无评论

Recognition of American sign language using modified deep residual CNN with modified canny edge segmentation

引用

Multimedia Tools and Applications 2025年 1-28页

作者： Hariharan, U. Devarajan, Devikanniga Kumar, P. Santhosh Rajkumar, K. Meena, M. Akilan, T. Department of Computer Science and Engineering Apex Institute of Technology Chandigarh University Punjab Mohali140413 India Department of Computer Science and Engineering GITAM University Karnataka Bengaluru561203 India Department of Information Technology Faculty of Engineering and technology SRM Institute of Science and Technology Ramapuram Tamil Nadu Chennai600089 India Department of Data Science and Business Systems School of Computing Faculty of Engineering and Technology SRM Institute of Science and Technology Kattankulathur Tamil Nadu Chennai603203 India Department of Computer Science and Engineering Sri Sairam Engineering College Tamil Nadu Tambaram602109 India Research Coordinator Dept. of IQAC Amity University Kolkata West Bengal 700135 India

American Sign Language (ASL) recognition aims to recognize hand gestures, and it is a crucial solution to communicating between the deaf community and hearing people. However, existing sign language recognition algorithms still have some drawbacks, such as difficulty recognizing hand movements and low recognition accuracy for most sign language recognition. A Modified Convolutional Neural Network (MCNN) deep residual 101 classifier-based American Sign Language identification system has been developed to address this problem. There are three main parts presents in the method. The first part is pre-processing the images to remove the noise, enhance the contrast of the picture, adjust the contrast level and smoothen the picture using various filters. The second part is the segmentation, and it’s used to partition the image using a modified canny edge detection method by removing all weak edges present in the image. Finally, classification will be done using the Modified CNN deep residual 101 classifiers. By classifying the image, the American Sign Language is accurately identified. This process is conducted through images. The outcome shows that the suggested approach has a 0.97% accuracy and a 0.05% False Positive Rate. Other CNN architectures such as resNet 50 and resNet 18 reached 0.90% and 0.80% accuracy, respectively, which is lower than our proposed method. In addition, the resNet 101 classifier effectively recognizes the difficult hand gestures through the image data. It obtains high recognition accuracy for 36 signs is 0 to 9 numbers and a to z alphabets from American Sign Language. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Optimizing Healthcare Accessibility and Effectiveness through the Implementation of AI Solutions for Remote Patient Monitoring 3

Optimizing Healthcare Accessibility and Effectiveness throug...

引用

3rd IEEE World Conference on Applied Intelligence and Computing, AIC 2024

作者： Singh, Jagendra Sahu, Dinesh Prasad Kumar, Ashwini Ahmad, Salman Khursheed Majumder, Jayeeta Gangopadhyay, Sourav Bennett University School of Computer Science Engineering & Technology Greater Noida India Camellia Institute of Technology & Management Department of Computer Science & Engineering Jaipur India Gl Bajaj Institute of Technology & Management Dept. of Applied Computational Science and Engineering Greater Noida India Brainware University Department of Hospital Management Kolkata India

ISBN: (纸本)9798350384598

This study investigates the optimization of healthcare accessibility and effectiveness by integrating artificial intelligence (AI) solutions within Internet of Things (IoT) systems for remote patient monitoring. Embracing a holistic approach, we deploy medical sensors, including temperature, pressure, heart rate, and oxygen sensors, to collect patients' continuous health metrics (vital signs). These sensed data are then transmitted to an Arduino controller, responsible for managing data communication with the cloud infrastructure using a Wi-Fi module that facilitates data storage and analysis. Stored data are accessed by authorized healthcare professionals, who exploit machine learning models, including Artificial Neural Networks (ANN), Logistic Regression (LR) and Decision Tree (DT), to forecast patients' health outcome in advance using the sensed data. A rigorous evaluation demonstrated the IoT system's effectiveness, based on testing with sensed data from ten patients over a month, where 70% of the data were used for training the models, and the rest for testing. The proposed IoT system demonstrates the success;ANN exhibited the highest accuracy at 97.88%, followed by LR at 93.43%, and DT at 91.20%. Precision, recall, and F1 scores further validated the models' ability to predict health outcome accurately. Evaluation of the accuracy and loss curves showed the convergence of the models towards optimal performance. This study contributes to remote healthcare monitoring and provides a promising IoT paradigm for realizing ubiquitous remote monitoring, optimized healthcare effectiveness, and patient outcome through AI-enabled IoT systems. © 2024 IEEE.

关键词： Remote patient monitoring

来源：评论

学校读者我要写书评

暂无评论

Improving Image Classification Using Attention Mechanisms in Deep Learning Models 4

Improving Image Classification Using Attention Mechanisms in...

引用

4th International Conference on Advancement in Electronics and Communication engineering, AECE 2024

作者： Sharma, Vinod Mishra, Atul Kumar Kumar, Anil Shakti, Nanda Sharma, Madhuri Srinju, M. Slp University Dept of Cse Gwalior India Madhav University Dept. of Commerce & Management Pindwara India Mangalayatan University Dept. of Eee Aligarh India Usha Martin University Dept. of Engineering and Applied Sciences Ranchi India Chri Campus University Dept. of Life Sciences Gwalior India Dhanalakshmi Srinivasan University Dept. of Ai and Data Science Coimbatore India

ISBN: (纸本)9798350364729

In this study, we explore the application of attention mechanisms to enhance deep learning models in the context of image classification. We assess several types of attention mechanisms across diverse image classification datasets. Our findings indicate substantial enhancements in model accuracy facilitated by these attention mechanisms. Additionally, we delve into implementation challenges such as computational overhead and the necessity for extensive datasets. Furthermore, we investigate the potential benefits of transfer learning and fine-tuning pretrained models integrated with attention mechanisms to optimize performance and streamline training efforts. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Enhancing Cloud Resource Allocation Efficiency with Harris Hawk Optimization Algorithm(HHOA)

Enhancing Cloud Resource Allocation Efficiency with Harris H...

引用

2024 IEEE International Students' Conference on Electrical, Electronics and Computer science, SCEECS 2024

作者： Manickavasagan, V. Petchiammal, Eswari Sudha, S Rahini Rajesh, C. Sree, S. Jindhiya Yamini, V.S. Sri Sai Ram Engineering College Dept. of Csbs Chennai India Velammal Institute of Technology Dept. of Information Technology Chennai India Panimalar Engineering College Dept. of Csbs Chennai India Vel Tech Multi Tech Dr.Rangarajan Dr.Sakunthala Engineering College Dept. of Artificial Intelligence and Data Science Chennai India Velammal Institute of Technology Department of Information Technology Chennai India

ISBN: (纸本)9798350348460

The demand for efficient cloud resource allocation has surged with the growing reliance on cloud computing for a wide array of applications. Optimizing resource allocation in cloud environments is crucial for cost-effectiveness, performance, and energy conservation. In this context, bio-inspired optimization algorithms have gained prominence for their ability to address complex allocation problems. This paper investigates the application of the Harris Hawk Optimization algorithm (HHOA) to enhance cloud resource allocation efficiency. HHOA, inspired by the hunting behavior of Harris's hawks, is an efficient and versatile metaheuristic algorithm known for its ability to find optimal solutions in complex, dynamic environments. By harnessing the HHOA algorithm, this study aims to improve the allocation of resources such as virtual machines and storage within cloud infrastructures as demanded by the tasks. In our proposed approach we present a comprehensive examination of HHOA's suitability for cloud resource allocation and compares its performance with conventional optimization algorithms like Stimulated Annealing (SA), Genetic Algorithm (GA) and Tabu Search (TS). Through a series of experiments, we found that Harris Hawk Optimization Algorithm (HHOA) is proven to be very effective in improving operational efficiency by enhanced reliability, lowered latency and lowered average response time. © 2024 IEEE.

关键词： Cloud computing GA HHOA Resource allocation Resource optimization Tabu search

来源：评论

学校读者我要写书评

暂无评论

Improved Dynamic Coupled Graph Convolutional Recurrent Networks for Traffic Forecasting

Improved Dynamic Coupled Graph Convolutional Recurrent Netwo...

引用

2023 IEEE International Conference on Big data, Bigdata 2023

作者： Han, Seunghoon Kim, Sungsoo Lim, Sungsu Chungnam National University Dept. of Computer Science & Engineering Daejeon Korea Republic of ETRI Smart Data Research Section Daejeon Korea Republic of

ISBN: (纸本)9798350324457

Traffic forecasting is a crucial application of the Intelligent Transportation System (ITS), with research focusing on various methods, from classical statistical approaches to graph-based methods integrated with RNN-based approaches to capture spatial and temporal correlations simultaneously. During the traffic data collection phase, the absence of vehicles on each road or sensor malfunctions can result in the collection of traffic time series data as zeros. However, storing such zero values makes accurate traffic prediction more challenging. To address this challenge, we present a novel model for improving traffic forecasting using graph convolutional recurrent neural networks. The proposed method is evaluated on two real-world public benchmark datasets and compared with six baseline models, showcasing its superior performance. © 2023 IEEE.

关键词： Graph Neural Networks Spatio-Temporal Time-Series Traffic Forecasting

来源：评论

学校读者我要写书评

暂无评论

Unified mechanism-specific amplification by subsampling and group privacy amplification 24

Unified mechanism-specific amplification by subsampling and ...

引用

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Jan Schuchardt Mihail Stoian Arthur Kosmala Stephan Günnemann Dept. of Computer Science & Munich Data Science Institute Technical University of Munich Dept. of Engineering University of Technology Nuremberg

ISBN: (纸本)9798331314385

Amplification by subsampling is one of the main primitives in machine learning with differential privacy (DP): Training a model on random batches instead of complete datasets results in stronger privacy. This is traditionally formalized via mechanism-agnostic subsampling guarantees that express the privacy parameters of a subsampled mechanism as a function of the original mechanism's privacy parameters. We propose the first general framework for deriving mechanism-specific guarantees, which leverage additional information beyond these parameters to more tightly characterize the subsampled mechanism's privacy. Such guarantees are of particular importance for privacy accounting, i.e., tracking privacy over multiple iterations. Overall, our framework based on conditional optimal transport lets us derive existing and novel guarantees for approximate DP, accounting with Rényi DP, and accounting with dominating pairs in a unified, principled manner. As an application, we analyze how subsampling affects the privacy of groups of multiple users. Our tight mechanism-specific bounds outperform tight mechanism-agnostic bounds and classic group privacy results.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-layer network modelling of genomic and transcriptomic data to investigate the response to checkpoint inhibitors in NSCLC

Multi-layer network modelling of genomic and transcriptomic ...

引用

2023 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2023

作者： Mascolo, Davide Farina, Lorenzo Petti, Manuela Sapienza University of Rome Program in Data Science Rome Italy Sapienza University of Rome Dept. of Computer Control and Management Engineering Rome Italy

ISBN: (纸本)9798350337488

Recent innovations and developments in molecular biology and biotechnology have made it possible to acquire and store large omics datasets. In particular genomics, transcriptomics and the study of their relationship represent the key elements to understand how genotype influences phenotype. In this work we investigate the potential of multi-layer network modeling in integrating genomic and transcriptomic data of 152 advanced non-small cell lung cancer (NSCLC) patients treated with anti-PD-(L)1 therapy. For the transcriptomic layer, we performed differential expression and differential co-expression analyses in order to identify a subset of key genes in differentiating responder patients from non-responders. Adding to the genomic-transcriptomic model other three layers related to immune, myeloid and curated immunotherapy-based literature signatures we obtained a 5-layer Patient Similarity Network. The application of Similarity Network Fusion algorithm revealed a statistically significant stratification of patients (p-value ≤ 0.05) which allows the identification of two clusters characterized by patients responding and not responding to immunotherapy. © 2023 IEEE.

关键词： immunotherapy multi-layer network Network-based omics integration non-small cell lung cancer patient Stratification

来源：评论

学校读者我要写书评

暂无评论

Performance Evaluation of data Imputation Methods for Graph Deep Learning-Based Traffic Prediction

Performance Evaluation of Data Imputation Methods for Graph ...

引用

2023 IEEE International Conference on Big data, Bigdata 2023

作者： Kim, Jeongseon Kim, Sungsoo Lim, Sungsu Chungnam National University Dept. of Computer Science & Engineering Daejeon Korea Republic of Smart Data Research Section Etri Daejeon Korea Republic of

ISBN: (纸本)9798350324457

We present five approaches for interpolating missing data in real-world traffic datasets using basic, statistical, and generative model methods. The effectiveness of these approaches is evaluated using graph deep learning models, and the results show that imputing missing data improves the performance of traffic prediction models, especially when dealing with a higher proportion of missing data. The experiments are conducted on two real-world datasets, METR-LA and PEMS-BAY, to evaluate the proposed methods comprehensively. The results demonstrate that imputing missing data significantly enhances the performance of traffic prediction models, particularly when dealing with a higher proportion of missing data. © 2023 IEEE.

关键词： Deep Learning on Graphs Missing data Imputation Spatio-Temporal Time-Series

来源：评论

学校读者我要写书评

暂无评论

Superpixel guided weighted low-rank representation for nonlinear hyperspectral unmixing 3

Superpixel guided weighted low-rank representation for nonli...

引用

3rd IEEE India Geoscience and Remote Sensing Symposium, InGARSS 2023

作者： Ahmad, Touseef Sahadevan, Anand S Raha, Soumyendu Lyngdoh, Rosly B Putrevu, Deepak Indian Space Research Organization Gujarat Ahmedabad380015 India Indian Institute of Science Dept. of Computational and Data Sciences Bangalore Karnataka560012 India Department of Space Meghalaya Umiam793103 India

ISBN: (纸本)9798350325591

The Generalized Bilinear Mixing model (GBM) has gained significant prominence in the field of nonlinear hyperspectral image unmixing, with recent advancements utilizing superpixel segmentation for the extraction of homogeneous patches. However, the application of the Low-Rank Representation (LRR) constraint to each superpixel has been observed to disproportionately diminish smaller singular values, resulting in associated challenges. Building upon our prior investigation of the weighted LRR (WLRR) approach within the robust GBM model, which effectively mitigated bias by assigning appropriate weights to singular values, we extend this framework in the current study. Here, we employ entropy rate superpixel (ERS) segmentation to extract homogeneous patches, applying the weighted low-rank constraint and developing the robust GBM model (RGM-SG-WLRR) for nonlinear unmixing analysis. The proposed model is effectively solved using the Alternating Direction Method of Multipliers (ADMM) iterative scheme. Through extensive experimental analyses conducted on both simulated and real data cubes, our findings highlight the superior performance of the RGM-SG-WLRR model over alternative methods, as evidenced by its notable improvements in terms of root-mean-square error and signal-to-reconstruction error. © 2023 IEEE.

关键词： Superpixels

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：