检索结果-内蒙古大学图书馆

2023 International Conference on Self Sustainable Artificial Intelligence Systems, ICSSAS 2023

作者： Agarwal, Pinky Gupta, Naman Bharadwaj, Yashita Yadav, Anju Mathur, Pratistha Manipal University Jaipur Computer Application Jaipur India Jaipur India Manipal University Jaipur Information Technology Jaipur India

ISBN: (纸本)9798350300857

Oral cancer is a malignancy that can manifest in multiple areas of the mouth, including the tongue, lips, oral mucosa, gums, floor of the mouth, and other soft tissues. Despite the development of advanced diagnostic and therapeutic methods, the incidence and prevalence of oral cancer continue to be a cause for concern. This condition can have devastating effects on patients, including significant physical and emotional distress. Detecting oral cancer at an early stage can improve a patient's chances of survival and reduce the need for aggressive treatment. In recent years, machine learning techniques have demonstrated promising results in the detection of oral cancer. This research aims to explore the role of vision transformers in oral cancer classification as benign and malignant. The results obtained from the vision transformer are compared with Machine learning techniques like SVM, extreme Gradient Boosting (XGBoost), and Random Forest. Overall, the results show that ViT outperformed SVM, Random Forest, and XGBoost in oral cancer detection and stage classification with an accuracy of 96.7%. The accuracy, precision, recall, and F1-score of ViT are higher than other methods. The findings of the research show that all four machine learning methods examined can produce high precision rates in detecting oral cancer, with the Vision Transformer technique achieving superior results as vision transformers are designed to capture global context information in images. The proposed system has great potential to advance the precision and efficiency of oral cancer diagnosis, which can lead to more prompt and effective treatment. © 2023 IEEE.

关键词： Diseases

来源：评论

学校读者我要写书评

暂无评论

Music Emotion Prediction based on Hybrid Approach Combining Lyrical and Acoustic Approaches 6

Music Emotion Prediction based on Hybrid Approach Combining ...

引用

6th International Conference on Intelligent Computing and Control Systems, ICICCS 2022

作者： Bisht, Shristi Kanakia, Harshil T. Thakur, Pallavi Sardar Patel Institute of Technology Master of Computer Application Department Mumbai India

ISBN: (纸本)9781665410359

With the advancements in multimedia technology in recent years, research on music is getting more significant and popular. It is necessary to study the relationship between music lyrics and the emotion of a user. Music based emotion analysis can help with the improvement of an individual's mood. This can be applied in the form of music recommendation systems, advertisement, etc. In this paper, we propose an emotion based prediction system using music lyrics (using NLP) and its acoustic features. Various moods of a single music are taken into consideration,for this reason, Russel's space is used to calculate its valence and arousal to define the emotion. The logistic regression is used for prediction and comparative study is done between various classification models for prediction and the best model is taken into consideration, separately for acoustic and lyrics based emotion prediction,Train/Test split is performed to get a 60:40 ratio for the respective approaches and the final emotion is predicted based on the hybrid approach. © 2022 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Research on Visible Light Equipment Image Data Enhancement technology Based on Generative Model

Research on Visible Light Equipment Image Data Enhancement T...

引用

International Conference on Big Data and Information Analytics (BigDIA)

作者： Youchen Wang Zhen Luan Zongkai Tian Huayi Zou Qi Liu Yuting Yang Beijing Institute of Computer Technology and Application Beijing China

ISBN: (数字)9798350354621

ISBN: (纸本)9798350354638

Visible light remote sensing imaging technology is an important tool for detecting and analyzing equipment. Massive image data is crucial for training and verifying automatic equipment recognition models. Although there are already corresponding image libraries, there are still problems such as insufficient image types and unsatisfactory quality. In order to solve these problems, this paper compares the use of several different GAN variants, including CGAN, WGAN-CP, WGAN-GP, and DDPM in the diffusion model to enhance the visible light remote sensing equipment image library. The effect is to expand the image data set by leveraging the advantages of each model, thereby supporting the implementation of more accurate and robust equipment identification in remote sensing applications.

关键词： Training Analytical models Image recognition Data enhancement Imaging Big Data Diffusion models Data models Libraries Remote sensing

来源：评论

学校读者我要写书评

暂无评论

A Brute-Force Attack Detection Method Based on the Combination of Artificial Neural Network and Auto Encoder

A Brute-Force Attack Detection Method Based on the Combinati...

引用

Cloud Computing, Big Data applications and Software Engineering (CBASE), International Conference on

作者： Xiaojun Ma Qingyuan Li Xiaofei Wang Quan Wen Yingming Zeng Beijing Institute of Computer Technology and Application Beijing China

ISBN: (数字)9798331506582

ISBN: (纸本)9798331506599

The ongoing advancement of information technology has led to the emergence of a multitude of network attack methodologies and techniques. As a prevalent form of network attack, it is imperative to undertake targeted research on the deliberate disruption of network traffic. Two issues currently impede the efficacy of machine learning and deep learning models in detecting brute force attacks. First, the presence of noise in the traffic data reduces the accuracy and robustness of detection. Second, when an attacker modifies the frequency, intensity, time, and other parameters of their attack to evade detection, the characteristics of the attack traffic become more complex and diverse. This, in turn, results in a notable decline in detection accuracy. In order to address the aforementioned issues, this paper presents a novel detection model that integrates the capabilities of an artificial neural network (ANN) and an auto-encoder (AE). The proposed model leverages the ANN's proficiency in recognizing intricate data patterns and the AE's effectiveness in removing noise, thereby facilitating precise detection of brute-force attacks. The experimental results demonstrate that the proposed detection model exhibits an enhanced accuracy rate of 99.93% in CICIDS-2017, exhibiting superior overall detection performance compared to existing algorithmic models such as AE and DAE. Additionally, it demonstrates a reduced false alarm rate and effective mitigation of the impact of noise and interference on attack data.

关键词： Time-frequency analysis Accuracy Prevention and mitigation Noise Telecommunication traffic Feature extraction Brute force attacks Robustness Pattern recognition Software engineering

来源：评论

学校读者我要写书评

暂无评论

Distributed Aperture Coherent Synthesis Method based on Two-stage Compensation 2

Distributed Aperture Coherent Synthesis Method based on Two-...

引用

2nd IEEE International Conference on Signal, Information and Data Processing, ICSIDP 2024

作者： Liu, Zhongwei Zhou, Dongming Xiang, Rongrong Chen, Jie Wu, Tianhang Ding, Tianyu Liu, Xiaochuan Gao, Hongwei Beihang University School of Electronic and Information Engineering Beijing China School of Information and Electronics Beijing Institute of Technology Radar Research Lab Beijing China Beijing Institute of Computer Technology and Application Department of Data Technology and Application Beijing China Beijing Institute of Radio Measurement Radar R&d Center Beijing China

ISBN: (纸本)9798331515669

To solve the problem that it is difficult to accurately estimate the coherent parameters of distributed aperture radar under the condition of low signal-to-noise ratio, a distributed coherent synthesis method based on two-stage compensation is proposed in this paper. This method utilizes prior information such as system layout, beam pointing, and inertial navigation station location to perform coarse alignment of echoes at the first stage in time, and then uses cross-correlation method to estimate coherent parameters of delay and phase within a small range, and perform precise compensation of time and phase at the second stage, so as to realize the coherent synthesis of multiple radar echoes. Experimental results show that the proposed method can effectively realize distributed coherent synthesis processing under low signal-to-noise ratio. © 2024 IEEE.

关键词： Signal to noise ratio

来源：评论

学校读者我要写书评

暂无评论

Radioactive Waste Repackaging Methods Based on Genetic Algorithms and Simulation for Reactor Decommissioning 5

Radioactive Waste Repackaging Methods Based on Genetic Algor...

引用

5th International Conference on computer, Big Data and Artificial Intelligence, ICCBD+AI 2024

作者： Wu, Tong Wang, Deying Chen, Fengjuan Zhang, Nan Yang, Fan Wang, Jiaqi Xu, Meng Computer Application Institute of Nuclear Industry Beijing China Beijing University of Technology Key Laboratory of Urban Security and Disaster Engineering of Ministry of Education Beijing China

ISBN: (纸本)9798331533991

After going through the lifecycle of design, construction, and operation, nu-clear reactors need go to decommissioning phase according to requirement of nuclear regulator or need of environmental protection. Decommissioning, as the last phase of a nuclear facility's lifecycle, is characterized by its difficulty, long duration, and high cost. During the decommissioning process, various types of waste are generated, such as solid, liquid, gas waste, and mixed waste. In order to protect the surrounding environment from radioactive contamination and prevent excessive radiation exposure to workers and residents, radioactive waste needs to be processed, prepared, transported, stored, and disposed of according to standards. In response to the low space utilization of steel containers due to the irregular shapes of waste when packing decommissioned reactor components, this paper optimizes the process using genetic algorithm to determine the optimal number of cut chunks for reactor components and the best placement of cut chunks in the steel container. First, the complex three-dimensional irregular part packing problem is projected onto a two-dimensional plane for study to simplify the problem's difficulty;second, genetic algorithm is used for individual selection, crossover, and mutation to obtain the optimal layout of individuals;finally, the optimal number of cut chunks in different areas is determined based on the optimization results of individuals, and a global packing solution for all cut chunks is obtained. The results indicate that the proposed method not only provides the optimal cutting solutions for different cutting areas of reactor components and the best placement of cut chunks in the steel container but also satisfies all cutting constraints, potentially providing insights for future reactor decommissioning work © 2024 IEEE.

关键词： Steel containers

来源：评论

学校读者我要写书评

暂无评论

A DoS Attack Detection System Based on CNN and BiLSTM for Internet of Vehicles

A DoS Attack Detection System Based on CNN and BiLSTM for In...

引用

Information Systems and Computing technology (ISCTech), International Conference on

作者： Bin Wang Ying Shang Fangming Wang Yingming Zeng Beijing Institute of Computer Technology and Application Beijing China

ISBN: (数字)9798350379860

ISBN: (纸本)9798350379877

As the Internet of Vehicles (IoV) becomes established and continues to develop, its reliability and security assurance is becoming increasingly important. Denial of Service (DoS) attacks, characterized by ease of implementation by attackers and a variety of attack methods, have become one of the most serious security threats facing inter-vehicle and intra-vehicle networks. In response to the demand for DoS attack detection in IoV environments, we propose a deep learning-based intrusion detection model, which integrates 1D convolutional neural network (1D-CNN) and bi-directional long short-term memory (BiLSTM). On the basis of preprocessing techniques, including data cleaning and imbalanced data resampling, our proposed model is able to capture spatial features of network traffic using 1D-CNN and learn temporal features of traffic using BiLSTM. Experimental results demonstrate that the proposed model performs superior detection performance than the other six classic techniques on two public datasets, CICIDS2017 and Car-Hacking.

关键词： Support vector machines Computational modeling Bidirectional long short term memory Intrusion detection Telecommunication traffic Feature extraction Convolutional neural networks Security Reliability Internet of Vehicles

来源：评论

学校读者我要写书评

暂无评论

Multi-stage Dental Visual Detection Based on YOLOv8: Dental 3D CBCT 2nd

Multi-stage Dental Visual Detection Based on YOLOv8: Dental...

引用

2nd International ToothFairy2: Multi-Structure Segmentation in CBCT Volumes, ToothFairy 2024, 3D Teeth Landmarks Detection Challenge, 3DTeethLand 2024, Semi-supervised Teeth Segmentation, STS 2024 held in conjunction with the 27th International Conference on Medical Image Computing and computer-Assisted Intervention, MICCAI 2024

作者： Zheng, Zhihao Ma, Dongdong He, Lian Cheng, Miao Liu, Jianhao Chengdu Institute of Computer Application Chinese Academy of Science Sichuan Chengdu610041 China School of Computer Science and Technology University of Chinese Academy of Science Beijing100049 China Shenzhen CBPM-KEXIN Banking Technology Company Limited Guangdong Shenzhen518206 China

ISBN: (纸本)9783031889769

With the popularity of panoramic X-ray imaging (PXI) and cone beam computed tomography (CBCT) in dental treatment planning and comprehensive prognostic assessment, the demand for effective tooth segmentation is growing. Although 2D panoramic X-rays can effectively detect invisible caries, impacted teeth and supernumerary teeth in children, 3D CBCT has more significant advantages in orthodontic and endodontic treatment, mainly due to its low radiation dose and accurate three-dimensional images. Although 3D images provide more detailed anatomical information, which helps accurate diagnosis and treatment planning, they also face challenges such as difficult data acquisition, radiation exposure, increased processing complexity and metal artifacts. In the field of tooth segmentation, although deep learning methods have made significant progress, the tooth segmentation task still faces many challenges. Compared with tooth semantic segmentation, tooth instance segmentation requires more refined boundary information and accurate capture of tooth morphology, which is particularly difficult in irregular gingival tissue and complex tooth morphology. In addition, the quality of tooth segmentation depends largely on detailed manual annotation, which is not only time-consuming and labor-intensive, but also prone to errors and subjectivity. Recently, STS MICCAI 2024 launched a 3D tooth segmentation challenge, which aims to explore tooth segmentation algorithms based on semi-supervised learning. By participating in this competition, this study explored semi-supervised learning methods through yolov8 [11], aiming to reduce dependence on a large amount of labeled data and train the model by utilizing a large amount of unlabeled data. Finally, among 77 teams, it achieved a score of 0.77 on the validation set, ranking 8th, and a score of 0.77 on the test set. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Intelligent Malware Detection Using Sentence Transformer and Hybrid Deep Learning Architecture 10

Intelligent Malware Detection Using Sentence Transformer and...

引用

10th International Conference on Signal Processing and Communication, ICSC 2025

作者： Kumar, Sanjeev Sapru, Yudhishthira Sapru, Sugandha Janet, B. Centre for Development of Advanced Computing Cyber Security Technology Division Mohali India Indian Computer Emergency Response Team CERT-In Delhi India Pggc Panjab University Department of Mathematics Chandigarh India National Institute of Technology Department of Computer Application Tiruchirappalli India

ISBN: (纸本)9798331511890

The complex evasion techniques adopted by the malware creator make signature-based malware detection approaches unable to detect unknown malware attacks. This study proposes a new malware detection and classification approach with significantly less dependence on manual feature engineering. The bidirectional encoder representations from transformers (BERT) are used to model the opcode sequences of the malware file as the semantic behavior of malware. After that, the vectorized feature vector is fed to the hybrid one-dimensional convolution neural network (1D-CNN) and the gated recurrent unit (GRU) model for feature reduction and normalization before training a Softmax classifier. A hybrid learning architecture is designed with a custom 1D-CNN and a GRU model to take the strengths of both types of networks. The proposed scheme is evaluated using the Microsoft BIG dataset, consisting of 1625 assembly files belonging to 9 distinct malware families. The proposed method produced 95.77% accuracy even with the smaller dataset, which we believe will boost performance by adding a more diverse dataset. It uses the next-generation BERT language model, which gives a promising future area for malware defenders. This research presents malware detection without much reliance on manual feature engineering or signature-based detection, making it suitable for detecting unknown malware attacks. © 2025 IEEE.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

PD-YOLO:Colon Polyp Detection Model Based on Enhanced Small-Target Feature Extraction

引用

computers, Materials & Continua 2025年第1期82卷 913-928页

作者： Yicong Yu Kaixin Lin Jiajun Hong Rong-Guei Tsai Yuanzhi Huang New Engineering Industry College Putian UniversityPutian351100China Engineering Research Center of Big Data Application in Private Health Medicine(Putian University) Fujian Province UniversityPutian351100China Computer College Guangdong University of Science&TechnologyDongguan523668China

In recent years,the number of patientswith colon disease has increased *** polyps are the precursor lesions of colon *** not diagnosed in time,they can easily develop into colon cancer,posing a serious threat to patients’lives and health.A colonoscopy is an important means of detecting colon ***,in polyp imaging,due to the large differences and diverse types of polyps in size,shape,color,etc.,traditional detection methods face the problem of high false positive rates,which creates problems for doctors during the diagnosis *** order to improve the accuracy and efficiency of colon polyp detection,this question proposes a network model suitable for colon polyp detection(PD-YOLO).This method introduces the self-attention mechanism CBAM(Convolutional Block Attention Module)in the backbone layer based on YOLOv7,allowing themodel to adaptively focus on key information and ignore the unimportant *** help themodel do a better job of polyp localization and bounding box regression,add the SPD-Conv(Symmetric Positive Definite Convolution)module to the neck layer and use deconvolution instead of *** results indicate that the PD-YOLO algorithm demonstrates strong robustness in colon polyp *** to the original YOLOv7,on the Kvasir-SEG dataset,PD-YOLO has shown an increase of 5.44 percentage points in AP@0.5,showcasing significant advantages over other mainstream methods.

关键词： Polyp detection YOLOv7 SPD-Conv CBAM deconvolution

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：