检索结果-内蒙古大学图书馆

VireNet-SSD: object detection model for resource-constrained applications based on self-organized operational neural networks

引用

Neural Computing and Applications 2025年第14期37卷 8547-8569页

作者： Kamath, Vidya Renuka, A. Department of Computer Science and Engineering Manipal Institute of Technology Manipal Academy of Higher Education Karnataka Manipal576104 India

Discovering deep learning-based computer vision solutions for use with constrained devices is exceptionally hard, and the trade-offs are often too undermining. Deep learning models are enormous, which makes it challenging to deploy them on constrained platforms. The convolutional neural network is the fundamental framework for majority of the models that are currently in use. However, operational neural networks have recently shown to be a better option to the convolutional equivalents on a variety of tasks due to their heterogeneous nature and greater resemblance to the functioning of biological neurons. The question of whether heterogeneous models could function on constrained devices and be deployed in real time remains a major concern. To address this problem, an object detection model architecture based on a single-shot multi-box detector with self-organized operational neural networks as its backbone was developed, which can perform efficiently on constrained devices such as Raspberry Pi. The resultant backbone architecture was named as VireNet. In contrast to homogeneous conventional deep learning networks that use convolutions, heterogeneous networks were chosen to develop VireNet, which provides a more productive and effective solution. Furthermore, an in-depth explanation of the design space has been provided to aid any future research that is associated with this architectural search. This new approach might mark the very beginning of the use of heterogeneity to address issues on devices with constrained resources. © The Author(s) 2025.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Soundtrack Success: Unveiling Song Popularity Patterns Using Machine Learning Implementation

引用

SN computer science 2024年第3期5卷 1-11页

作者： Arora, Shruti Rani, Rinkle Department of Computer Science and Engineering Thapar Institute of Engineering and Technology Punjab Patiala India

Music streaming services are getting increasingly popular as a result of the prevalent need for web and smart gadgets. Melophiles are drawn toward a range of musical genres and create a unique digital footprint. The emotional reactions of the music listeners may cause physiological changes as well as neurological benefits. However, any musician should avoid being influenced by an algorithmic recommendation presented in the research. The most prominent aspect for any musician is to stay true to their own musical style and to be novel, honest, and sincere. The main objective of this research work is to provide precise music predictions to the artists before they launch their music in the market. This work also aims to improve the accuracy of the prediction models using some feature-engineering-based techniques to manipulate the Spotify dataset and predict whether the song will be a hit or not. The proposed work leverages state-of-the-art machine learning models, including Linear Regression, Lasso Regression, Ridge Regression, Elastic Net, Random Forest, GBM, and neural networks to explore the predictive factors influencing the inclusion of a particular music track in the featured Spotify Hit-50 list. Furthermore, it offers descriptive statistical analysis of the features that can help in popularity prediction. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd 2024.

关键词： Hit song science audio features Music information retrieval Music streaming data Regression

来源：评论

学校读者我要写书评

暂无评论

Proposed Solution for Currency Classification and Fake Currency Detection Using Machine Learning 4

Proposed Solution for Currency Classification and Fake Curre...

引用

4th International Conference on Sentiment Analysis and Deep Learning, ICSADL 2025

作者： Narale, Dnyaneshwari. B. Pujar, Anita. M. Pardeshi, K.R. Walchand Institute of Technology Computer Science and Engineering Solapur India

ISBN: (纸本)9798331523923

The rapid advancement of digital technologies has greatly impacted financial systems, creating challenges in accurately classifying currencies and detecting counterfeits in various real-world scenarios. Traditional approaches, such as manual inspections and specialized tools like UV light detectors, tend to be time-consuming, expensive, and less effective against sophisticated counterfeiting methods. Recent developments in machine learning and image processing have shown potential, utilizing supervised machine learning models. Nonetheless, issues remain in managing variations due to wear, folds, inconsistent lighting, and different currency designs. This research introduces a robust machine learning-based system for classifying currencies and detecting counterfeit bills. By combining advanced image processing techniques to extract essential features (like texture, edges, and color histograms) with a hybrid model that integrates CNN, Random Forest, and XGBoost, the proposed framework tackles challenges related to real-world variability and subtle signs of counterfeiting. The goal of the system is to achieve high accuracy and reliability in identifying counterfeit notes while ensuring it can scale and adapt to new counterfeiting methods. © 2025 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Challenge Response Protocol using Multivariate Polynomial Cryptography

Challenge Response Protocol using Multivariate Polynomial Cr...

引用

2025 IEEE International Students' Conference on Electrical, Electronics and computer science, SCEECS 2025

作者： Jaiswal, Chandra Keshwar Sharma, Arpit Chaturvedi, Atul Allenhouse Institute of Technology Computer Science and Engineering Kanpur India

ISBN: (纸本)9798331529833

The objective of this paper is to define the meaning of a multivariate public key using two or more multivariate encrypted parties A and B which targets to improve the existing multivariate public key encryption. Two polynomials are created by a central authority for the protocol's setup phase. The client generates a public key using a secret key, then uses his secret key to respond to a challenge the server issues and submits it to the server. The response is hashed to ensure its security. Any design, including ones with physically unclonable functions, can use this multivariate polynomial cryptography scheme since it is appropriate for post-quantum computing and quantum resistant encryption. This protocol, which protects the CCA technique of post-quantum communications, validated the promise of multivariate polynomials for information security. The work makes an effort to portray polynomial implementation and polynomial contribution as security-sensitive jobs. © 2025 IEEE.

关键词： Quantum computers

来源：评论

学校读者我要写书评

暂无评论

Advancements in Isogeny-Based Cryptography: A Mathematical Approach to Post-Quantum Security

Advancements in Isogeny-Based Cryptography: A Mathematical A...

引用

2025 IEEE International Students' Conference on Electrical, Electronics and computer science, SCEECS 2025

作者： Soni, Abhay Singh, Sunidhi Chaturvedi, Atul Allenhouse Institute of Technology Computer Science and Engineering Kanpur India

ISBN: (纸本)9798331529833

In this paper, we have discussed isogeny-based cryptography as a rather appealing option for post-quantum security. We provide analysis of current protocols like SIDH (Supersingular Isogeny Diffie-Hellman) and SIKE (Supersingular Isogeny Key Encapsulation) setting up fundamental issues with their security and performance. In this work we introduce a novel isogeny pathfinding algorithm that employs graph theory as well as probabilistic models in order to lower its computational complexity. The results show significant improvement in both efficiency and security, strengthening the case for the use of isogeny based systems in a quantum-resilient cryptographic space. © 2025 IEEE.

关键词： Encryption algorithms

来源：评论

学校读者我要写书评

暂无评论

Analysis of community question-answering issues via machine learning and deep learning:State-of-the-art review

引用

CAAI Transactions on Intelligence technology 2023年第1期8卷 95-117页

作者： Pradeep Kumar Roy Sunil Saumya Jyoti Prakash Singh Snehasish Banerjee Adnan Gutub Department of Computer Science and Engineering Indian Institute of Information TechnologySuratIndia Indian Institute of Information Technology DharwadIndia Department of Computer Science and Engineering National Institute of TechnologyPatnaIndia The York Management School University of YorkYorkUK Computer Engineering Department College of Computer&Information SystemsUmm Al-Qura UniversityMakkahSaudi Arabia

Over the last couple of decades,community question-answering sites(CQAs)have been a topic of much academic *** have often leveraged traditional machine learning(ML)and deep learning(DL)to explore the ever-growing volume of content that CQAs *** clarify the current state of the CQA literature that has used ML and DL,this paper reports a systematic literature *** goal is to summarise and synthesise the major themes of CQA research related to(i)questions,(ii)answers and(iii)*** final review included 133 *** research themes include question quality,answer quality,and expert *** terms of dataset,some of the most widely studied platforms include Yahoo!Answers,Stack Exchange and Stack *** scope of most articles was confined to just one platform with few cross-platform *** with ML outnumber those with ***,the use of DL in CQA research is on an upward trajectory.A number of research directions are proposed.

关键词： answer quality community question answering deep learning expert user machine learning question quality

来源：评论

学校读者我要写书评

暂无评论

Performance Evaluation of Biometric Authentication Using Fragment Jaya Optimizer-Based Deep CNN with Multi-kernel SVM

引用

SN computer science 2024年第4期5卷 1-9页

作者： Umasankari, N. Muthukumar, B. Shanmuganathan, C. Department of Computer Science and Engineering Sathyabama Institute of Science and Technology Chennai India School of Computing Department of CSE Sathyabama Institute of Science and Technology Chennai India Rajalakshmi Engineering College Chennai India Department of Computer Science and Engineering SRM Institute of Science and Technology Ramapuram Chennai India

The earlier research clearly indicated that the bimodal authentication system has more efficiency than unimodal and multimodal. This is due to the reason for the best intact biometric traits of fingerprint and retina. There is a chance to additionally improve further performance of the proposed biometric trait combination by additional or alternate algorithms or methodologies. Therefore, in this research work, the multi-kernel support vector machine (MK-SVM), a machine learning classification approach, is proposed and is used for the implementation. In addition, a hybrid algorithm of fragment Jaya optimizer-based deep convolutional neural network (FJO-DCNN) approach is also used to improve the performance value for bimodal biometric authentication and classification. The recognition systems analyze both biometrics independently, and their conclusions are combined to determine whether to give or refuse access to the user in the end. According to the findings of the implementation, this work demonstrates more dependability than the cascaded biometric system. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd 2024.

关键词： Bimodal biometric authentication Deep convolutional neural network Fragment Jaya optimizer convolutional neural network Multi-kernel support vector machine

来源：评论

学校读者我要写书评

暂无评论

Comprehensive insights into healthcare IoT: the role of machine learning and deep learning approaches

引用

Multimedia Tools and Applications 2025年 1-42页

作者： Kumar, Atul Dewang, Rupesh Kumar Computer Science and Engineering Department Motilal Nehru National Institute of Technology Allahabad Uttar Pradesh Prayagraj211004 India

Conventional healthcare systems have long struggled to address the varied needs of large patient populations, often leading to inefficiencies and less-than-optimal outcomes. Yet, the advent of machine learning (ML) and deep learning (DL) models has heralded a transformative shift toward value-based care, enabling healthcare providers to offer personalized and remarkably effective treatments. Contemporary medical equipment and devices are equipped with internal applications that gather and store extensive patient data, forming a valuable resource for ML-powered predictive *** have conducted a systematic literature review (SLR) which mainly focuses on Theoretical background, challenges, Application framework, Organization of IoT in healthcare and HIoT *** the SLR process some efficient methodologies has been used. In the SLR process we have conducted Quality metadata analysis (QMA) which provides the complete statistical analysis by considering different *** learning algorithms enable the analysis of the extensive data produced by IoT devices, uncovering patterns and irregularities that may escape human detection. Deep learning, leveraging sophisticated neural network models, enhances the accuracy of medical image interpretation, disease prognosis, and forecasting patient outcomes. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Diagnosis

来源：评论

学校读者我要写书评

暂无评论

A Deep Learning Approach for Intelligent Diagnosis of Lung Diseases

引用

SN computer science 2024年第8期5卷 1-9页

作者： Paswan, Jai Dev Bhatia, Tarunpreet Lamba, Sonu Computer Science and Engineering Thapar Institute of Engineering and Technology Punjab Patiala 147001 India

According to the World Health Organization (WHO), lung diseases contribute to millions of fatalities globally each year. Pneumonia stands out as a leading cause, claiming the lives of approximately 2.5 million individuals. Tuberculosis accounted for 1.4 million deaths in 2019 alone. Furthermore, COVID-19, the coronavirus, has resulted in over 9 million reported fatalities worldwide as of July 2024. Addressing the burden of lung diseases is crucial for public health, emphasizing the necessity of initiatives aimed at prevention and management to enhance global health outcomes. Deep learning for medical image analysis serves as a valuable resource for both academic and industrial researchers keen on delving into medical image analysis. This study explores a range of image pre-processing techniques, including Gaussian blur, total variation denoising, histogram equalization, and CLAHE, applied to a chest X-ray image dataset. These techniques aim to enhance image quality, thereby facilitating more accurate disease classification by the model. The study focuses on multiclass classification en-compassing normal, viral pneumonia, bacterial pneumonia, tuberculosis, and COVID-19 classes. To accomplish this, four deep learning models-VGG16, VGG19, ResNet50, and DenseNet121 are employed for disease detection from chest X-ray images, with their performances compared. Results indicate that VGG16 achieves the highest accuracy, with 94.04% for training and 87.80% for testing. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Chest X-ray Deep learning models Image pre-processing Lung disease

来源：评论

学校读者我要写书评

暂无评论

Crop weed separation through image-level segmentation: an ensemble of modified U-Net and encoder–decoder

引用

Neural Computing and Applications 2025年第16期37卷 9887-9910页

作者： Ganapathy, Sannasi Srinivasan, Srinitish Bhopal462 002 India School of Computer Science and Engineering Vellore Institute of Technology Chennai600 127 India

Crop weed segmentation is one of the most challenging tasks in the field of computer vision. This is because, unlike other object detection or segmentation tasks, crop and weed are similar in terms of spectral features, shape, dimensions, etc. For precision agriculture to flourish in terms of smart spraying of crops, efficient systems to distinguish between crop and weed are the need of the hour, which if precise, will take a huge step toward solving the issue of food scarcity. To tackle this issue, we propose new ensemble architecture of two models—a U-Net with a modified backbone and an encoder–decoder. These networks learn to distinguish between soil and crop and soil and weed, respectively, whose ensemble gives state-of-the-art results on pixel-wise annotations of combined crop and weed images. Moreover, it also learns that the model captures un-annotated features since each component of the architecture learnt either crop or weed features to high precision. Finally, the proposed architecture is compared with the U-Net and SegNet, which are popular segmentation networks, and consistently achieved better results. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.

关键词： Semantic 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：