检索结果-内蒙古大学图书馆

2nd International Conference on Applied Algorithms, ICAA 2025

作者： Raj, Pronit Kumar, Chandrashekhar Shekhar, Harshit Kumar, Amit Paul, Kritibas Jana, Debasish Department of Computer Science and Engineering Heritage Institute of Technology Kolkata India

ISBN: (纸本)9783031845420

In today’s digital world, streaming platforms offer a vast array of movies, making it hard for users to find content matching their preferences. This paper explores integrating real-time data from popular movie websites using advanced HTML scraping techniques and APIs. It also incorporates a recommendation system trained on a static Kaggle dataset, enhancing the relevance and freshness of suggestions. By combining content-based filtering, collaborative filtering, and a hybrid model, we create a system that utilizes both historical and real-time data for more personalized suggestions. Our methodology shows that incorporating dynamic data not only boosts user satisfaction but also aligns recommendations with current viewing trends. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Data reduction

来源：评论

学校读者我要写书评

暂无评论

A new hybrid MobileNet-MLP model for secured data communication using signature verification and deep learning

引用

Iran Journal of computer Science 2024年第1期8卷 233-252页

作者： Pandiyaraju, V. Ganapathy, Sannasi Thangaramya, K. Nahar, R. Raunak Kannan, A. School of Computer Science and Engineering Vellore Institute of Technology Chennai India Department Computer Science and Engineering Education National Institute of Technical Teachers Training and Research (NITTTR) Bhopal India School of Computer Science and Engineering Vellore Institute of Technology Vellore India

Signature verification plays a critical role in various industries, including finance and document authentication. However, traditional verification techniques have limitations, such as a lack of robustness and an inability to effectively capture intricate patterns and relationships within the signature images. This paper introduces an innovative approach that addresses the drawbacks by integrating conventional machine learning methods with deep learning techniques. For this purpose, we design a new hybrid MobileNet model using deep learning to extract features from the signature images. Here, the Multilayer Perceptron (MLP) classifier is also used to carry out the binary classification over the extracted features. The experimental results of the proposed model demonstrate that the amalgamated technique surpasses the conventional methods and underscores the possibilities of merging traditional machine learning and deep learning in signature verification. Moreover, the proposed model overcomes the limitations of the traditional signature verification techniques by leveraging the strengths of both deep learning and traditional machine learning. Finally, it obtains a more robust and efficient solution and proves it by providing high security to the data in terms of accuracy. © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2024.

关键词： Convolutional neural network Feature extraction MobileNet Multilayer perceptron Signature verification

来源：评论

学校读者我要写书评

暂无评论

Evaluating CUDA-Aware Approximate Computing Techniques 3

Evaluating CUDA-Aware Approximate Computing Techniques

引用

3rd International Workshop on Resource AWareness of Systems and Society, RAW 2024

作者： Öz, Işıl Computer Engineering Department Izmir Institute of Technology Izmir Turkey

Approximate computing techniques offer performance improvements by performing inexact computations. Moreover, CUDA programs written to be executed on GPU devices employ specific features to utilize the parallel computation units of heterogeneous GPU architectures. While generic software-level approximate computing techniques have been applied to heterogeneous CUDA programs, CUDA-specific approaches may introduce promising performance improvements by not corrupting the target computations. In this work, we propose software approximation techniques for CUDA programs: kernel-aware loop perforation, partition-level synchronization, block-level atomic operations, and warp divergence elimination. We perform source code transformations on target benchmark programs by applying our techniques. We evaluate performance improvements by trading off accuracy in our target computations. Our experimental results reveal that CUDA-aware approximation techniques offer significant performance improvements at the expense of acceptable accuracy loss. © 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).

关键词： Graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

Optimizing NOMA Transmissions to Advance Federated Learning in Vehicular Networks

Optimizing NOMA Transmissions to Advance Federated Learning ...

引用

2024 IEEE Global Communications Conference, GLOBECOM 2024

作者： Chen, Ziru Ni, Zhou Guan, Peiyuan Wang, Lu Cai, Lin X. Hashemi, Morteza Li, Zongzhi Illinois Institute of Technology Department of Electrical and Computer Engineering Chicago United States University of Kansas Department of Electrical Engineering and Computer Science Lawrence United States University of Oslo Department of Informatics Oslo Norway Illinois Institute of Technology Department of Civil Engineering Chicago United States

ISBN: (纸本)9798350351255

Diverse critical data, such as location information and driving patterns, can be collected by IoT devices in vehicular networks to improve driving experiences and road safety. However, drivers are often reluctant to share their data due to privacy concerns. The Federated Vehicular Network (FVN) is a promising technology that tackles these concerns by transmitting model parameters instead of raw data, thereby protecting the privacy of drivers. Nevertheless, the performance of Federated Learning (FL) in a vehicular network depends on the joining ratio, which is restricted by the limited available wireless resources. To address these challenges, this paper proposes to apply Non-Orthogonal Multiple Access (NOMA) to improve the joining ratio in a FVN. Specifically, a vehicle selection and transmission power control algorithm is developed to exploit the power domain differences in the received signal to ensure the maximum number of vehicles capable of joining the FVN. Our simulation results demonstrate that the proposed NOMA-based strategy increases the joining ratio and significantly enhances the performance of the FVN. Index Terms - Federated Vehicular Network, NOMA © 2024 IEEE.

关键词： Data privacy

来源：评论

学校读者我要写书评

暂无评论

Machine Learning Models for Forest Fire Detection 5

Machine Learning Models for Forest Fire Detection

引用

5th International Conference on Data Intelligence and Cognitive informatics, ICDICI 2024

作者： Kulkarni, Radhika V. Kamble, Ruchita Joshi, Soham Jaiswal, Priten Jumbad, Varad Vishwakarma Institute of Technology Department of Computer Engineering Pune India

ISBN: (纸本)9798350389609

One of the most serious natural disasters endangering forest resources and areas is forest fires. Decreased casualties and enhanced firefighting depend on precise and timely forest fire detection. Due to their reliance on ground inspection and field-of-view limitations, conventional firefighting techniques are inadequate for monitoring vast areas. Unmanned aerial vehicles (UAVs) are very difficult to identify forest fires, isolate fire pixels from the backdrop using Red Green Blue (RGB), grayscale color models, and extract luminance and chrominance from the original image. Over the past several years, machine-learning methods have been used to analyze forest images for detecting fires like Convolutional Neural Networks (CNNs), Support Vector Machines (SVMs), Random Forests, and many more. The goal is to determine the model for this crucial task by examining performance metrics across various algorithms. This study compares models like CNN, U-Net, and conventional machine learning models based on their accuracy, precision, and F1 score. Through thorough testing and analysis, this study aims to fully assess each model's strengths and flaws, ultimately aiding to development of more dependable and efficient methods for detecting forest fires. © 2024 IEEE.

关键词： Unmanned aerial vehicles (UAV)

来源：评论

学校读者我要写书评

暂无评论

Self-supervised few-shot medical image segmentation with spatial transformations

引用

Neural Computing and Applications 2024年第30期36卷 18675-18691页

作者： Titoriya, Ankit Kumar Singh, Maheshwari Prasad Singh, Amit Kumar Department of Computer Science & amp Engineering National Institute of Technology Patna Ashok Rajpath Bihar Patna800005 India

Deep learning-based segmentation models often struggle to achieve optimal performance when encountering new, unseen semantic classes. Their effectiveness hinges on vast amounts of annotated data and high computational resources for training. However, a promising solution to mitigate these challenges is the adoption of few-shot segmentation (FSS) networks, which can train models with reduced annotated data. The inherent complexity of medical images limits the applicability of FSS in medical imaging, despite its potential. Recent advancements in self-supervised label-efficient FSS models have demonstrated remarkable efficacy in medical image segmentation tasks. This paper presents a novel FSS architecture that enhances segmentation accuracy by utilising fewer features than existing methodologies. Additionally, this paper proposes a novel self-supervised learning approach that utilises supervoxel and augmented superpixel images to further enhance segmentation accuracy. This paper assesses the efficacy of the proposed model on two different datasets: abdominal magnetic resonance imaging (MRI) and cardiac MRI. The proposed model achieves a mean dice score and mean intersection over union of 81.62% and 70.38% for abdominal images, and 79.38% and 65.23% for cardiac images. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Workout Pose Correction System using KNIFT 6

Workout Pose Correction System using KNIFT

引用

6th International Conference on Cybernetics, Cognition and Machine Learning Applications, ICCCMLA 2024

作者： Uke, Shailaja Ghote, Varad Mundhe, Ved Thorat, Shreyas Patil, Ujjwal Vishwakarma Institute of Technology Department of Computer Engineering Pune India

ISBN: (纸本)9798331505790

This paper proposes a system to assist users in performing exercise workouts by implementing MediaPipe Keypoint Neural Invariant Feature Transform (KNIFT). Exercise requires proper directions to follow as improper movements during exercise can lead to injuries. Mediapipe KNIFT offers real time tracking of the user's movement as compared to only pose detection. Using the coordinates of the joints, the angle between joints and speed of movement can be calculated which can be used to guide to user to follow proper exercise movements. Pygame library was used to make a user friendly interface, along with voice recognition to enable hands free usage. © 2024 IEEE.

关键词： Speech recognition

来源：评论

学校读者我要写书评

暂无评论

NRASV: Noise Robust ASV System for Audio Replay Attack Detection 5th

NRASV: Noise Robust ASV System for Audio Replay Attack Detec...

引用

5th International Conference on Deep Learning, Artificial Intelligence and Robotics, ICDLAIR 2023

作者： Chakravarty, Nidhi Dua, Mohit Department of Computer Engineering National Institute of Technology Kurukshetra India

ISBN: (纸本)9783031609343

Recent advancements in researching countermeasure systems against audio replay attacks have notably bolstered the resilience of Automatic Speaker Verification (ASV) Systems. However, current countermeasures face challenges in adapting to unknown attack variations. This limitation stems from the failure to capture context-specific details from the speech data, a crucial factor leading to the suboptimal performance of these systems against novel threats. To address this, our proposed anti-replay system integrates Gammatone features with t-distributed Stochastic Neighbor Embedding (t-SNE) for dimensionality reduction. In the classification phase, we employ diverse machine learning algorithms including Support Vector Machine (SVM), Random Forest (RF), K Nearest Neighbor (KNN), and Gradient Boost (Gboost). Our evaluation, conducted on the VSDC dataset in both controlled and noisy environments, introduces babble noise at varying levels (15dB, 10dB, 5dB, and 0dB). Remarkably, our model, GTCC-t-SNE with SVM, attains exceptional performance with Equal Error Rates (EER) of 2.2%, 3.3%, 3.8%, and 4.4% for the respective noise levels, demonstrating its robustness against audio replay attacks. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Dimensionality reduction

来源：评论

学校读者我要写书评

暂无评论

Blockchain-Driven Academic Learning Record Management in Higher Education: A Comprehensive Review of Methodologies, Applications, Benefits, and Challenges

引用

SN computer Science 2025年第5期6卷 1-33页

作者： Chaudhari, Sheetal Shirole, Mahesh Computer Science and Engineering Department Sardar Patel Institute of Technology Andheri Maharashtra Mumbai 400058 India Computer Engineering and Information Technology Department Veermata Jijabai Technological Institute Matunga Maharashtra Mumbai 400031 India

Blockchain technology has the potential to revolutionize the education sector by improving the legitimacy and security of academic procedures. The academic learning record management records the learner’s academic qualifications and achievements. Keeping track of higher education academic learning records is essential to assess students’ credentials and advance their careers. The COVID-19 outbreak highlighted the need for flexible and collaborative learning systems. However, the processes of recording credit accumulation, verification, and transfer are complicated by the disparate credit schemes found in different universities. This creates the need for an academic credit management system. Secure learning record management is necessary because validating the professional and academic accomplishments listed on the resumé currently poses a challenge. Blockchain can remedy this scenario by guaranteeing data privacy, reducing the possibility of hacking, and offering transparent, traceable academic credentials as well as verification of other achievements. Despite these benefits, challenges are still associated with implementing blockchain in academic learning record management. These include handling multiple credentials with a unique identity, absence of smart contracts, on-chain and centralized storage, scalability problems, transaction costs, restricted control over student data, and inadequate student data recording standards. The main objective of this review is to examine recent studies on blockchain-driven academic learning record management by categorizing them into learning record management and academic credit management. This paper identifies research gaps and suggests further studies and practical applications to enhance transparency, efficiency, and trust in the education system. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Academic bank of credits Academic credit management Blockchain Education Learning record management NEP 2020

来源：评论

学校读者我要写书评

暂无评论

CamoVeil: Digital Camouflage Pattern Generation System 1

引用

2nd International Conference on Advances in Data-driven Computing and Intelligent Systems, ADCIS 2023

作者： Ganganwar, Vaishali Rajput, Parshant Bhavana, N. Sinha, Swapnil Vaid, Aashna Department of Computer Engineering Army Institute of Technology Pune India

ISBN: (数字)9789819995318

ISBN: (纸本)9789819995301

Camouflage is the art of concealing. It is widely used to conceal your own self and your equipment from the enemy. It had been in nature since before humans, animals have used it to adapt to their surroundings. Camouflage is rooted in nature, and can be seen in animals to conceal them from predators. Its widespread application for military tactics only started with British zoologists. Camouflage now has come to play a crucial role in various domains with applications in military, hunting, and wildlife observations. The traditional methods of generating camouflage patterns that are generated manually are time-consuming and pose significant challenges in achieving optimal cover for different environments. In order to solve the issues and limitations, we in this paper have proposed an automated method to generate camouflage patterns tailored for specific terrains. We utilize image detection, and color extraction and produce pixelated patterns. The entire process involves segmentation, object detection, dominant color extraction, and pattern generation using K-means++. The CamoGen program is then employed to generate camouflage patterns by covering the canvas with polygons of respective colors and pixelating them into block-like patterns. The proposed method offers flexibility by allowing the generation of multiple patterns through parameter adjustments. Validation techniques are also integrated to ensure the desired areas are concealed. By digitizing and automating the camouflage pattern generation process, our model represents a significant advancement in the field and has potential applications in stealth monitoring and military exercises. Experimental results demonstrate the efficiency and effectiveness of our approach, particularly in natural environments such as forests, grasslands, and deserts. Further enhancements are required for designing camouflage patterns in artificial environments like urban settings. By digitizing and automating the camouflage pattern gen

关键词： Military applications

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：