检索结果-内蒙古大学图书馆

Deep Learning-based Environmental Sound Classification Using Feature Fusion and Data Enhancement

computers, Materials & Continua 2023年第1期74卷 1069-1091页

作者： Rashid Jahangir Muhammad Asif Nauman Roobaea Alroobaea Jasem Almotiri Muhammad Mohsin Malik Sabah M.Alzahrani Department of Computer Science COMSATS University IslamabadVehari CampusPakistan Department of Computer Science University of Engineering and Technology LahorePakistan Department of Computer Science College of Computers and Information TechnologyTaif UniversityTaif21974Saudi Arabia

Environmental sound classification(ESC)involves the process of distinguishing an audio stream associated with numerous environmental *** common aspects such as the framework difference,overlapping of different sound events,and the presence of various sound sources during recording make the ESC task much more complicated and *** research is to propose a deep learning model to improve the recognition rate of environmental sounds and reduce the model training time under limited computation *** this research,the performance of transformer and convolutional neural networks(CNN)are *** audio features,chromagram,Mel-spectrogram,tonnetz,Mel-Frequency Cepstral Coefficients(MFCCs),delta MFCCs,delta-delta MFCCs and spectral contrast,are extracted fromtheUrbanSound8K,ESC-50,and ESC-10,***,this research also employed three data enhancement methods,namely,white noise,pitch tuning,and time stretch to reduce the risk of overfitting issue due to the limited audio *** evaluation of various experiments demonstrates that the best performance was achieved by the proposed transformer model using seven audio features on enhanced *** UrbanSound8K,ESC-50,and ESC-10,the highest attained accuracies are 0.98,0.94,and 0.97 *** experimental results reveal that the proposed technique can achieve the best performance for ESC problems.

关键词： Environmental sound classification convolutional neural network deep learning transformer data augmentation

来源：评论

学校读者我要写书评

暂无评论

Neural network-guided motion blur estimation and patch representation method for face hallucination

引用

Neural Computing and Applications 2025年 1-28页

作者： Kumar, Banti Rajput, Shyam Singh Department of Computer Science and Engineering National Institute of Technology Patna Bihar Patna India

Face hallucination (or super-resolution) is a technique of generating high-resolution face images from low-resolution (LR) inputs. The current face hallucination algorithms struggle with motion blur, a common issue in captured images due to various factors, including camera defocusing, objects in motion, etc. To address this problem, a new motion blur robust face hallucination algorithm via neural network-guided motion blur estimation and patch representation (NNMEPR) is proposed in this paper. The NNMEPR algorithm first estimates the motion blur kernels i.e., length (λ) and angle (α) from a motion-blurred LR test faces using a neural network fitting approach with scale conjugate gradient (SCG) learning algorithm. Then, the estimated factor is embedded in the LR dictionary face images to make them compatible with the test images and mitigate the impact of motion blur from the reconstruction process. Moreover, the proposed algorithm employs a neighboring position patch representation in the correlation coefficient calculation process to sustain the sharp edges and texture details in the resulting HR face images. Experiments on standard datasets and locally captured real-world faces demonstrate the superior performance of the proposed NNMEPR algorithm over the existing state-of-the-art techniques. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.

关键词： Motion estimation

来源：评论

学校读者我要写书评

暂无评论

Q-Learning-Based Pesticide Contamination Prediction in Vegetables and Fruits

引用

computer Systems science & engineering 2023年第4期45卷 715-736页

作者： Kandasamy Sellamuthu Vishnu Kumar Kaliappan Department of Computer Science and Engineering KPR Institute of Engineering and TechnologyCoimbatore641407Tamil NaduIndia

Pesticides have become more necessary in modern agricultural ***,these pesticides have an unforeseeable long-term impact on people's wellbeing as well as the *** to a shortage of basic pesticide exposure awareness,farmers typically utilize pesticides extremely close to *** residues within foods,particularly fruits as well as veggies,are a significant issue among farmers,merchants,and particularly *** residual concentrations were far lower than these maximal allowable limits,with only a few surpassing the restrictions for such pesticides in *** is an obligation to provide a warning about this amount of pesticide use in *** technologies failed to forecast the large number of pesticides that were dangerous to people,necessitating the development of improved detection and early warning systems.A novel methodology for verifying the status and evaluating the level of pesticides in regularly consumed veggies as well as fruits has been identified,named as the Hybrid Chronic Multi-Residual Framework(HCMF),in which the harmful level of used pesticide residues has been predicted for contamination in agro products using Q-Learning based Recurrent Neural Network and the predicted contamination levels have been analyzed using Complex Event Processing(CEP)by processing given spatial and sequential *** analysis results are used to minimize and effectively use pesticides in the agricultural field and also ensure the safety of farmers and ***,the technique is carried out in a Python environment,with the results showing that the proposed model has a 98.57%accuracy and a training loss of 0.30.

关键词： Pesticide contamination complex event processing recurrent neural network Q learning multi residual level and contamination level

来源：评论

学校读者我要写书评

暂无评论

SMART DRIVING LICENSING SYSTEM AND AUTHORIZATION OF AUTONOMOUS VEHICLE INTEGRATING WIRELESS SENSOR NETWORKS AND IoT

引用

Telecommunications and Radio engineering (English translation of Elektrosvyaz and Radiotekhnika) 2024年第9期83卷 69-85页

作者： Santhakumar, D. Sudhan, M.B. Prasanna, R. Manikandan, S.P. Prathaban, Banu Priya Kanna, Rajesh R. Saravanan, G. Department of Computer Science and Engineering Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Chennai India Department of Computer Science and Engineering SRM Institute of Science and Technology Chennai Ramapuram India Department of Electronics and Communication Engineering SRM Institute of Science and Technology Kattankulathur Chennai India Department of Electronics and Communication Engineering Sree Sastha Institute of Engineering and Technology Chennai India Department of Networking and Communications School of Computing SRM Institute of Science and Technology Kattankulathur Chennai India Department of Electronics and Communication Engineering Sri Sairam Institute of Technology Tamil Nadu Chennai India

The advent of autonomous vehicles has revolutionized the automotive industry, offering promising advancements in safety, efficiency, and mobility. To integrate these autonomous vehicles into our society seamlessly, it is imperative to establish a robust licensing system that ensures the safe operation of automated vehicles on public roads. This article presents an effective smart Driving Licensing System (SDLS), a digital framework designed to regulate the existing licensing system and authorize licensing for autonomous vehicles. Through a comprehensive assessment process, the system assesses the drivers and the autonomous vehicle’s ability to navigate various road conditions, respond to dynamic environments, and prioritize safety for passengers and other road users. SDLS integrates high-level security through facial and fingerprint verification for identity verification to preserve licensing data. Hacking and unauthorized access to autonomous vehicles is prevented by robust cybersecurity protocols. The key benefits of SDLS include automated data collection, integration of wireless sensor networks (WSN) and Internet of Things (IoT), smart application processing, real-time license checks, cybersecurity measures, automated testing, and digital certification. The proposed system offers a comprehensive approach to safely integrating autonomous vehicles into our transportation ecosystem. The autonomous vehicles are integrated with antennas at rooftops or into bumpers, depending upon the vehicle design which can be connected with the SDLS for signal communication. Furthermore, the SDLS establishes a framework for continuous monitoring and updates, enabling ongoing assessments to keep pace with evolving technology and safety standards. © 2024 by Begell House, Inc.

关键词： Autonomous vehicles

来源：评论

学校读者我要写书评

暂无评论

MFR-CNN: A modified faster R-CNN approach based on bounding box and reliable score for cloth image retrieval

引用

Multimedia Tools and Applications 2024年 1-29页

作者： Kaur, Navneet Pandey, Shreelekha Kalra, Nidhi Department of Computer Science and Engineering Thapar Institute of Engineering & ampTechnology Punjab Patiala147004 India School of Computer Science and Engineering RV University South Karnataka Bangalore560059 India

Today, machine learning is used in a broad variety of applications. Convolution neural networks (CNN), in particular, are widely used to analyze visual data. The fashion industry is catching up to the growing usage of machine learning in a variety of fields throughout the world. The proposed model named MFR-CNN, which is a modified Faster R-CNN approach, has been presented in this manuscript. The effectiveness of an algorithm is validated on cloth attributes by established datasets such as the Deep Fashion dataset and the Fashion Landmark dataset, alongside a new dataset created for the same manually. The proposed model detects clothes using bounding boxes and further classifies them. Apart from this, retrieval functionality has also been added. The effectiveness of the proposed model is evident through the learning rate analysis done for the modules present in the same. On the metrics of accuracy, execution time, and total loss, the model proposed here is contrast with existing techniques named R-CNN, Fast R-CNN, and Faster R-CNN. It is found that MFR-CNN is 50% more accurate in contrast to R-CNN. Also, the accuracy of MFR-CNN is 30% and 7% more accurate than the Fast R-CNN and Faster R-CNN, respectively. Moreover, MFR-CNN is quite effective in contrast to other approaches as it takes less time for execution. It consumes 33 seconds less in contrast to R-CNN. Also, the time taken by the MFR-CNN is 17 seconds and 3 seconds less in comparison to Fast R-CNN and Faster R-CNN, respectively. From the experimental results it is being inferred that the MFR-CNN approach for detecting and classifying clothes is both efficient and effective. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Image retrieval

来源：评论

学校读者我要写书评

暂无评论

Conditional adversarial segmentation and deep learning approach for skin lesion sub-typing from dermoscopic images

引用

Neural Computing and Applications 2024年第26期36卷 16445-16463页

作者： Mirunalini, P. Desingu, Karthik Aswatha, S. Deepika, R. Deepika, V. Jaisakthi, S.M. Department of Computer Science Engineering Sri Sivasubramaniya Nadar College of Engineering Chennai603110 India School of Computer Science and Engineering Vellore Institute of Technology Chennai Campus Chennai600127 India

Automatic skin lesion subtyping is a crucial step for diagnosing and treating skin cancer and acts as a first level diagnostic aid for medical experts. Although, in general, deep learning is very effective in image processing tasks, there are notable areas of the processing pipeline in the dermoscopic image regime that can benefit from refinement. Our work identifies two such areas for improvement. First, most benchmark dermoscopic datasets for skin cancers and lesions are highly imbalanced due to the relative rarity and commonality in the occurrence of specific lesion types. Deep learning methods tend to exhibit biased performance in favor of the majority classes with such datasets, leading to poor generalization. Second, dermoscopic images can be associated with irrelevant information in the form of skin color, hair, veins, etc.;hence, limiting the information available to a neural network by retaining only relevant portions of an input image has been successful in prompting the network towards learning task-relevant features and thereby improving its performance. Hence, this research work augments the skin lesion characterization pipeline in the following ways. First, it balances the dataset to overcome sample size biases. Two balancing methods, synthetic minority oversampling TEchnique (SMOTE) and Reweighting, are applied, compared, and analyzed. Second, a lesion segmentation stage is introduced before classification, in addition to a preprocessing stage, to retain only the region of interest. A baseline segmentation approach based on Bi-Directional ConvLSTM U-Net is improved using conditional adversarial training for enhanced segmentation performance. Finally, the classification stage is implemented using EfficientNets, where the B2 variant is used to benchmark and choose between the balancing and segmentation techniques, and the architecture is then scaled through to B7 to analyze the performance boost in lesion classification. From these experiments, we find

关键词： Pipelines

来源：评论

学校读者我要写书评

暂无评论

Classifying the state of knowledge-based question answering: patterns, progress, and prospects

引用

International Journal of computers and Applications 2025年第1期47卷 93-105页

作者： Sharma, Sumit Jain, Sarika Tiwari, Mukesh Kumar lal, Sohan Department of Computer Science and Engineering Quantum University Roorkee Uttarakhand Roorkee India Department of Computer Applications National Institute of Technology Haryana Kurukshetra India

In response to inquiries posed in natural languages, question-answering systems (QASs) produce responses. The capabilities of early QASs are limited because they were designed for certain domains. The current generation of QASs focuses on the types of inquiries that users typically enquire, the features of the data sources that are used, and the formats in which the right responses are produced. Many QASs have been developed since the 1960s when research on the subject first started. There are several cutting-edge frameworks for answering questions that can be used to categorize the QA system according to a number of criteria. The necessity for an extensive survey on QASs comes naturally in order to determine the future focus of research in this field. In this study, QASs are analyzed by the number of criteria used to categorize question-answering (QA) and provide the dimension using a variety of standards to cover all the QAS. We describe the current state of research in each QAS area and make recommendations for its future direction. Based on the dimension criteria outlined in the research methodology, research studies are chosen for additional review and comparison. Constructed some criteria to classify the QASs, we provide the classification and go over the benefits and drawbacks of each class's QAS, as well as the problems with the corresponding study. Results are detailed summarized and compared. As a result of this study, a general architecture of frameworks for QA over knowledge base (KB) is drawn in addition to the comparative analysis of systems, serving as a reference for academics working on related topics. © 2024 Informa UK Limited, trading as Taylor & Francis Group.

关键词： Question answering

来源：评论

学校读者我要写书评

暂无评论

Improved Set Algebra-Based Heuristic Technique for Training Multiplicative Functional Link Artificial Neural Networks for Financial Time Series Forecasting

引用

SN computer science 2024年第5期5卷 567页

作者： Behera, Sudersan Kumar, AVS Pavan Nayak, Sarat Chandra Department of Computer Science and Engineering School of Engineering & amp Technology GIET University Gunupur India Department of Computer Science and Engineering School of Technology GITAM University Hyderabad India

The current study is defined by two main aims. An effective strategy for improving local search is to combine the Set Algebra-Based Heuristic Algorithm (SAHA) algorithm with the Nelder-Mead simplex method. The approach outlined above, referred to as the Improved SAHA (ISAHA), and has the ability to produce superior outcomes. The multiplicative functional link artificial neural network (MFLANN) is an improved version of the functional link artificial neural network that promotes exploration by replacing the summing unit in the output layer with a multiplication unit. Moreover, the combination of ISAHA and MFLANN results in the development of ISAHA-MFLANN, a sophisticated hybrid forecasting model. The main assessment of the hybrid model rests on its ability to forecast complex and dynamic financial time series. It's possible to get around the problems that come with traditional learning-based MFLANN techniques by using MFLANN's advanced approximation features and ISAHA's resilient global search features together. Experimental verification using two stock market datasets and three currency exchange rates demonstrates the validity of the idea. The results show that the suggested improved SAHA hybrid learning works well at improving six standard benchmark functions. The ISAHA-MFLANN model is also statistically significant at accurately capturing the volatility that is inherent in the financial time series. In addition, it surpasses other models such as SAHA-MFLANN, Monarch Butterfly Optimization-MFLANN, Particle Swarm Optimization-MFLANN, Genetic Algorithm-MFLANN, Gradient Descent-MFLANN, Support Vector Machine (SVM), Multilayer Perceptron (MLP), and Auto Regressive Integrated Moving Average (ARIMA). © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Exchange rate forecasting Improved Set Algebra-Based Heuristic Algorithm Multiplicative functional link artificial neural network Stock market forecasting

来源：评论

学校读者我要写书评

暂无评论

Adaptive Neighborhood Adjustment Strategy Based On MOHHO and NSGA-III Algorithms for Feature Selection

引用

IAENG International Journal of Applied Mathematics 2024年第5期54卷 917-935页

作者： Papasani, Anusha Durgam, Revathi Devarakonda, Nagaraju Computer Science and Engineering Department Koneru Lakshmaiah Education Foundation AndhraPradesh Vaddeswaram522502 India Computer Science and Engineering Department AVN Institute of Engineering and Technology Telangana Hyderabad501510 India Science and Engineering Department VIT-AP University AP Amaravati522237 India

The efficiency of multi-objective evolutionary algorithms (MOEAs) in tackling issues with multiple objectives is examined. However, it is noted that current MOEA-based feature selection techniques often converge towards the center of the Pareto front due to inadequate selection forces. The study proposes the utilization of a novel approach known as MOEA/D, which partitions complex multi-objective problems into smaller, more feasible single-objective sub-problems. Each sub-problem may then be addressed using an equal amount of computational resources. The predetermined size of the neighborhood used by MOEA/D may lead to a delay in the algorithm's merging and reduce the effectiveness of the failure. The paper proposes the Adaptive Neighbourhood Adjustment Strategy (ANAS) as a novel approach to improve the efficiency of multi-objective optimisation algorithms in order to tackle this issue. The ANAS algorithm allows for adaptive adjustment of the subproblem neighborhood size, hence enhancing the trade-off between merging and variety. In the following section of the study, a novel feature selection technique called MOGHHNS3/D-ANA is introduced. This technique utilizes ANAS to expand the potential solutions for a particular subproblem. The approach evaluates the chosen features using the Regulated Extreme Learning Machine (RELM) classifier on sixteen benchmark datasets. The experimental results demonstrate that MOGHHNS3/D-ANA outperforms four commonly employed multi-objective techniques in terms of accuracy, precision, recall, F1 score, coverage, hamming loss, ranking loss, and training time, error. The APBI approach in decomposition-based multi-objective optimization focuses on handling constraints by adjusting penalty parameters to guide the search towards feasible solutions. On the other hand, the ANA approach focuses on dynamically adjusting the neighborhood size or search direction based on the proximity of solutions in the detached space to adapt the search process.

关键词： Multiobjective optimization

来源：评论

学校读者我要写书评

暂无评论

A lightweight approach for intrusion detection in WSNs based on DCGAN

引用

International Journal of Information technology (Singapore) 2025年第2期17卷 951-957页

作者： Devi, Manu Nandal, Priyanka Sehrawat, Harkesh Department of Computer Science and Engineering University Institute of Engineering and Technology Maharshi Dayanand University Haryana India Department of Computer Science and Engineering Maharaja Surajmal Institute of Technology New Delhi India

Ensuring strong security procedures is crucial in the rapidly advancing realm of wireless sensor networks (WSNs) in order to protect sensitive data and preserve network integrity. The resource limitations and unpredictable environments that characterize WSNs frequently provide challenges for traditional intrusion detection systems (IDS). To overcome these issues, this research introduces a unique lightweight intrusion detection system that utilizes Deep Convolutional Generative Adversarial Networks (DCGAN). The proposed system employs wireless sensor network blackhole flooding selective forwarding (WSN-BFSF) dataset for training and evaluation. The suggested approach improves the IDS’s detection capabilities in addition to reducing computing overhead. The findings show that the lightweight DCGAN-based IDS provides an improved defense against new security threats, making it a good fit for real-time deployment in WSNs. Here, the performance metrics like receiver operating characteristic (ROC) is evaluated to check the efficiency of the model. The proposed model is assessed with the existing studies based on metrics such as accuracy, precision, recall and f1-score. The results show that the proposed approach attains the highest accuracy (94%) thus outperforming the existing methods and studies. © Bharati Vidyapeeth's Institute of computer Applications and Management 2024.

关键词： Deep Learning (DL) Generative Adversarial Networks (GAN’s) Intrusion Detection System (IDS) Lightweight DCGAN Wireless Sensor Networks (WSN’s)

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：