检索结果-内蒙古大学图书馆

FDM-RTDETR: A Multi-Scale Small Target Detection Algorithm

IEEE Access 2025年 13卷 88747-88761页

作者： Wang, Hongya Yu, Yongtao Tang, Zhaoxia Huaiyin Institute of Technology Faculty of Computer and Software Engineering Jiangsu Huaian223003 China

To address challenges in uncrewed aerial vehicles (UAV) object detection including complex backgrounds, severe occlusion, dense small objects, and varying lighting conditions, we propose FDM-DETR, a novel detection algorithm specifically designed for small objects in UAV imagery. This method effectively captures global image information by fusing multi-scale spatial features and performing feature extraction in the frequency domain within the backbone network. We design a Dynamic Feature Interaction (DIFI) module with position-based biases in the encoder, enhancing the model’s perception of local features for small objects. In the neck network, we introduce a Multi-Scale Feature Enhancement Pyramid (MSFEP) module to improve feature extraction capabilities for small object detection. Compared to RT-DETR, our improved model achieves performance gains of 2.5% and 2.7% in AP on the Vis-Drone2019 validation and test sets, respectively. While maintaining low computational complexity and parameter count, the method demonstrates significant improvements in detection performance. FDM-DETR exhibits robust practicality and reliability in UAV-based small object detection tasks. © 2013 IEEE.

关键词： Target drones

来源：评论

学校读者我要写书评

暂无评论

software Test Data Management Based on Knowledge Graph

引用

Informatica (Slovenia) 2024年第16期48卷 27-36页

作者： Gao, Li Qiu, Junlin Chen, Guanhua Faculty of Computer and Software Engineering Huaiyin Institute of Technology Huai’an223003 China

As software development models and methods mature, large-scale software systems emerge. However, a critical challenge remains: the lack of a comprehensive software test data management model that integrates basic data management with advanced knowledge reasoning. To address this issue, we developed a software test data management model based on knowledge graphs, enabling intelligent management and reasoning of software test data. The model incorporates an entity extraction model based on a feed-forward neural network, a knowledge graph integration method based on graph databases, and a knowledge reasoning submodule based on deep learning. To validate the effectiveness of our model, we evaluated the performance of each component individually. Our deep learning-based entity extraction model achieved an accuracy of 0.92, a recall of 0.88, and an F1 score of 0.90, significantly outperforming traditional methods such as regular expressions and dictionary-based approaches. Utilizing Cypher for graph database querying, our system provides accurate answers with a response time of 0.12 seconds, outperforming SQL and SPARQL-based querying methods. Furthermore, our approach excels in knowledge-based reasoning with an accuracy of 0.89 and site coverage of 0.81, surpassing both ontology-based and graph-based reasoning methods. These results highlight the enhanced construction, querying, and reasoning capabilities of our knowledge graph-based approach for managing software testing data. © 2024 Slovene Society Informatika. All rights reserved.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

An Experimental Study on Dynamic Lifelong Learning with GPT for Mitigating Catastrophic Forgetting in Aspect-Based Sentiment Analysis

引用

IEEE Access 2025年 13卷 90316-90332页

作者： Huang, Huang Mustafa, Mumtaz Begum Asemi, Adeleh University of Malaya Faculty of Computer Science and Information Technology Department of Software Engineering Kuala Lumpur50603 Malaysia Branch Faculty of Engineering Department of Computer Engineering Isfahan81551-39998 Iran

GPT is widely recognized as one of the most versatile and powerful large language models, excelling across diverse domains. However, its significant computational demands often render it economically unfeasible for individuals and small businesses, underscoring the need for efficient, domain-specific alternatives. In Aspect-Based Sentiment Analysis (ABSA), existing models are typically optimized for single domains, facing challenges in performing effectively across multiple domains. A key issue, known as "catastrophic forgetting,"arises when models trained on one domain lose previously learned knowledge upon exposure to new domain data. This leads to two significant problems: limited cross-domain generalization and difficulty in retaining prior knowledge while learning domain-specific information. To address these challenges, we introduce the Dynamic Lifelong Learning Aspect-Based Sentiment Analysis GPT (DllaGPT), a model designed to handle multiple domains while mitigating catastrophic forgetting. Leveraging datasets from four ABSA domains - Laptops, Restaurants, Tweets, and Finance - this study fine-tunes a pretrained GPT model from HuggingFace sequentially across domains. DllaGPT employs a mechanism to retain real data from earlier domains during new domain training, effectively preserving prior knowledge. Experimental results highlight that DllaGPT achieves an average accuracy of 0.85 and a Backward Transfer (BWT) score of -0.09 across the four domains, showcasing its high accuracy and robust lifelong learning capabilities. © 2013 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Deepfake Audio Detection for Urdu Language Using Deep Neural Networks

引用

IEEE Access 2025年 13卷 97765-97778页

作者： Ahmad, Omair Khan, Muhammad Sohail Jan, Salman Khan, Inayat University of Engineering and Technology Department of Computer Software Engineering Mardan Pakistan Arab Open University Faculty of Computer Studies A’Ali732 Bahrain University of Engineering and Technology Department of Computer Science Mardan Pakistan

Audio Deepfakes, which are highly realistic fake audio recordings driven by AI tools that clone human voices, With Advancements in Text-Based Speech Generation (TTS) and Vocal Conversion (VC) technologies have enabled it easier to create realistic synthetic and imitative speech, making audio Deepfakes a common and potentially dangerous form of deception. Well-known people, like politicians and celebrities, are often targeted. They get tricked into saying controversial things in fake recordings, causing trouble on social media. Even kids’ voices are cloned to scam parents into ransom payments, etc. Therefore, developing effective algorithms to distinguish Deepfake audio from real audio is critical to preventing such frauds. Various Machine learning (ML) and Deep learning (DL) techniques have been created to identify audio Deepfakes. However, most of these solutions are trained on datasets in English, Portuguese, French, and Spanish, expressing concerns regarding their correctness for other languages. The main goal of the research presented in this paper is to evaluate the effectiveness of deep learning neural networks in detecting audio Deepfakes in the Urdu language. Since there’s no suitable dataset of Urdu audio available for this purpose, we created our own dataset (URFV) utilizing both genuine and fake audio recordings. The Urdu Original/real audio recordings were gathered from random youtube podcasts and generated as Deepfake audios using the RVC model. Our dataset has three versions with clips of 5, 10, and 15 seconds. We have built various deep learning neural networks like (RNN+LSTM, CNN+attention, TCN, CNN+RNN) to detect Deepfake audio made through imitation or synthetic techniques. The proposed approach extracts Mel-Frequency-Cepstral-Coefficients (MFCC) features from the audios in the dataset. When tested and evaluated, Our models’ accuracy across datasets was noteworthy. 97.78% (5s), 98.89% (10s), and 98.33% (15s) were remarkable results for the RNN+LSTM

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

A Package Level Regression Test Selection Approach for Java software Systems

引用

IEEE Access 2025年 13卷 87848-87861页

作者： Al-Refai, Mohammed Hammad, Mahmoud M. Jordan University of Science and Technology Faculty of Computer and Information Technology Computer Science Department Irbid22110 Jordan Faculty of Engineering and Information Technology Ajman United Arab Emirates Jordan University of Science and Technology Faculty of Computer and Information Technology Software Engineering Department Irbid22110 Jordan

Regression testing of software systems is an important and critical activity yet expensive and resource-intensive. An approach to enhance its efficiency is Regression Test Selection (RTS), which selectively re-executes a subset of relevant tests that are impacted by code modifications. Previous studies on static and dynamic RTS for Java software have shown that selecting tests at the class level is more effective than using finer granularities like methods or statements. Nevertheless, RTS at the package level, which is a coarser granularity than class level, has not been thoroughly investigated or evaluated for Java projects. To address this gap, we propose PKRTS, a static package-level RTS approach that utilizes the structural dependencies of the software system under test to construct a package-level dependency graph. PKRTS analyzes dependencies in the graph and identifies relevant tests that can reach modified packages, i.e., packages containing altered classes. In contrast to conventional static RTS techniques, PKRTS implicitly considers dynamic dependencies, such as Java reflection and virtual method calls, among classes belonging to the same package by treating all those classes as a single cohesive node in the dependency graph. We evaluated PKRTS on 885 revisions of 9 open-source Java projects, with its performance compared to Ekstazi, a state-of-the-art dynamic class-level approach, and STARTS, a state-of-the-art static class-level approach. We used Ekstazi as the baseline to measure the safety and precision violations of PKRTS and STARTS. The results indicated that PKRTS outperformed static class-level RTS in terms of safety violation, which measures the extent to which relevant test cases are missed. PKRTS showed an average safety violation of 2.29% compared to 5.94% safety violation of STARTS. Despite this, PKRTS demonstrated lower precision violation and lower reduction in test suite size than class-level RTS, as it selects higher number of irrelevant te

关键词： Java programming language

来源：评论

学校读者我要写书评

暂无评论

HybridGAD: Identification of AI-Generated Radiology Abstracts Based on a Novel Hybrid Model with Attention Mechanism

引用

computers, Materials & Continua 2024年第8期80卷 3351-3377页

作者： TugbaÇelikten Aytug Onan Department of Software Engineering Faculty of TechnologyManisa Celal Bayar UniversityManisa45140Turkey Department of Computer Engineering Faculty of Engineering and ArchitectureIzmir KatipÇelebi UniversityIzmir35620Turkey

Class Title:Radiological imaging method a comprehensive overview *** GPT paper provides an overview of the different forms of radiological imaging and the potential diagnosis capabilities they offer as well as recent advances in the *** and Methods:This paper provides an overview of conventional radiography digital radiography panoramic radiography computed tomography and cone-beam computed *** recent advances in radiological imaging are discussed such as imaging diagnosis and modern computer-aided diagnosis ***:This paper details the differences between the imaging techniques the benefits of each and the current advances in the field to aid in the diagnosis of medical ***:Radiological imaging is an extremely important tool in modern medicine to assist in medical *** work provides an overview of the types of imaging techniques used the recent advances made and their potential applications.

关键词： Generative artificial intelligence AI-generated text detection attention mechanism hybrid model for text classification

来源：评论

学校读者我要写书评

暂无评论

A Fast Algorithm for Computing the Deficiency Number of a Mahjong Hand

引用

Chinese Journal of Electronics 2024年第6期33卷 1383-1398页

作者： Xueqing YAN Yongming LI Sanjiang LI School of Computer Science Shaanxi Normal University School of Mathematics and Statistics Shaanxi Normal University Centre for Quantum Software and Information Faculty of Engineering and Information TechnologyUniversity of Technology Sydney

The tile-based multiplayer game Mahjong is widely played in Asia and has also become increasingly popular worldwide. Face-to-face or online, each player begins with a hand of 13 tiles and players draw and discard tiles in turn until they complete a winning hand. An important notion in Mahjong is the deficiency number(*** number in Japanese Mahjong) of a hand, which estimates how many tile changes are necessary to complete the hand into a winning hand. The deficiency number plays an essential role in major decision-making tasks such as selecting a tile to discard. This paper proposes a fast algorithm for computing the deficiency number of a Mahjong hand. Compared with the baseline algorithm, the new algorithm is usually 100 times faster and, more importantly,respects the agent's knowledge about available tiles. The algorithm can be used as a basic procedure in all Mahjong variants by both rule-based and machine learning-based Mahjong AI.

关键词： Machine learning algorithms Knowledge based systems Decision making Asia Games

来源：评论

学校读者我要写书评

暂无评论

A mathematical multi-objective optimization model and metaheuristic algorithm for effective advertising in the social internet of things

引用

Neural Computing and Applications 2025年第6期37卷 4797-4821页

作者： Molaei, Reza Rahsepar Fard, Kheirollah Bouyer, Asgarali Department of Computer Engineering and Information Technology University of Qom Qom Iran Department of Software Engineering Azarbaijan Shahid Madani University Tabriz Iran Department of Software Engineering Faculty of Engineering and Natural Science Istinye University Istanbul Turkey

The integration of social networks with the Internet of Things (IoT) has been explored in recent research, giving rise to the Social Internet of Things (SIoT). One promising application of SIoT is viral marketing, which plays a critical role in modern advertising. However, current solutions proposed for viral marketing are often impractical in real-world scenarios, and there has been limited investigation into viral marketing within the context of SIoT. To address this issue, we present a multi-objective integer linear programming model named ILP-ASoT that facilitates advertising through SIoT. Our approach introduces two realistic parameters: the node sociality rate and the node appropriateness rate as an advertising destination. By balancing the cost of selecting a seed set with the number of users who receive the advertisement, our model offers an appropriate trade-off. Since this is proven as an NP-hard problem, A PSO-based metaheuristic algorithm is proposed to provide near-optimal solutions for advertising in SIoT, called PSO-ASoT. Through simulations, we show that our PSO-ASoT algorithm outperforms state-of-the-art algorithms in several metrics, such as influence spread, advertising cost, and duplicated advertising rate. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Heuristic algorithms

来源：评论

学校读者我要写书评

暂无评论

PoolKANNeXt: A new pooling-based Kolmogorov Arnold convolutional neural network

引用

Alexandria engineering Journal 2025年 128卷 144-152页

作者： Lv, Fangxing Wei, Qing Huang, Yuwen Tuncer, Turker Dogan, Sengul Özyurt, Fatih School of Computer Heze University Heze274015 China Department of Digital Forensics Engineering Technology Faculty Firat University Elazig Turkey Department of Software Engineering Engineering Faculty Firat University Elazig Turkey

The Kolmogorov–Arnold Network (KAN) is a new-generation neural network. It provides an alternative to multilayer perceptrons (MLPs). PoolFormer showed that pooling alone can mix features efficiently. We propose PoolKANNeXt, a CNN that merges the KAN structure with pooling-based feature mixing. This design targets high accuracy with a low number of the learnable parameters. We evaluated PoolKANNeXt on six image datasets, five biomedical and one general (CIFAR-10). The model comprises four stages. In the stem stage, a ConvNeXt-style patchify block converts each 224 × 224 × 3 image into a 56 × 56 × 96 tensor. In the main stage, average pooling first mixes local features;then two parallel 3 × 3 convolutions—one followed by GELU and the other by Swish—extract complementary representations, and a 1 × 1 convolution scales the combined output and adds it back to the input via a residual connection. In the downsampling stage, strided 3 × 3 convolutions halve spatial dimensions and double the channel count. In the output stage, global average pooling produces a feature vector that feeds into a softmax classifier. PoolKANNeXt achieved over 90 % accuracy on all datasets and reached 99.43 % on CIFAR-10, ranking among the top five models on that benchmark. PoolKANNeXt offers a lightweight yet powerful architecture. Its innovative combination of pooling and dual-activation KAN blocks yields strong performance across diverse tasks. The design is scalable and adaptable to larger or more complex datasets. © 2025 The Authors

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

COVID-19 Prediction Model Empowered with Fused Computational Intelligence Technique

引用

IEIE Transactions on Smart Processing and Computing 2025年第1期14卷 109-117页

作者： Khan, Muhammad Adnan Naseer, Iftikhar Ali, Muhammad Nadeem Kim, Byung-Seo Department of Software Faculty of Artificial Intelligence & Software Gachon University Seongnam13120 Korea Republic of Faculty of Computer Science & Information Technology The Superior University Lahore54600 Pakistan Department of Software and Communications Engineering Hongik University Sejong Campus Korea Republic of

The novel Coronavirus (COVID-19) spread rapidly around the world and caused overwhelming effects on the health and economy of the world. It first appeared in Wuhan city of China and was declared a pandemic by the World Health Organization (WHO). Many researchers, as well as experts in clinical and artificial intelligence experts, are working together to control the rapid spread of COVID-19 with early detection. This study focused on intelligent prediction for coronavirus using computational intelligence approaches (IPC-FCIA) like convolutional neural networks, support vector machines, and fuzzy logic techniques. The proposed IPC-FCIA model is based on two sections namely the training section and the validation section. Features fusion and decision-level fusion are used in this study to enhance the performance of the recommended IPC-FCIA model. The proposed model predicts the early detection of COVID-19 in two types COVID-negative and COVID-positive. The benchmark results of the model show an accuracy of 97.66 % on decision-level fusion in the detection of COVID-19. The proposed model can be helpful for medical experts as well as COVID-19-affected patients. © 2025 Institute of Electronics Engineers of Korea. All rights reserved.

关键词： COVID-19

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：