检索结果-内蒙古大学图书馆

computer Vision and Machine Intelligence (CVMI), International Conference on

作者： M Sivabalamurugan T R Swapna Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Coimbatore India

ISBN: (数字)9798350376876

ISBN: (纸本)9798350376883

The rise of deepfake technology presents a critical challenge to the integrity of digital media, prompting the need for advanced forgery detection techniques. This paper proposes a novel approach for detecting deepfake image forgery by integrating local surface geometry analysis into the detection process. Leveraging the inherent alterations in local surface geometry caused by deepfake generation, our method extracts and merges this information with the original image, enhancing the discriminative features essential for accurate identification of manipulated content. Using classification networks such as Resnet, Densenet, and Efficient, our method achieves a remarkable accuracy rate of above $98 \%$ in correctly identifying deepfake images for the classification models. This high accuracy underscores the effectiveness of our approach in detecting sophisticated deepfake content. By combining local surface geometry analysis with deep learning-based classification, our method represents a significant advancement in the field of deepfake detection, offering a promising solution to combat the proliferation of manipulated media and preserve the authenticity of digital imagery.

关键词： Geometry Deepfakes Accuracy Media Feature extraction Forgery Robustness Data mining Surface treatment Resilience

来源：评论

学校读者我要写书评

暂无评论

Efficient Underwater Object Detection Using Deep Neural Networks

Efficient Underwater Object Detection Using Deep Neural Netw...

引用

computing, Power and Communication Technologies (IC2PCT), IEEE International Conference on

作者： Shaik Atheeq Rasool R Annamalai B. Natarajan Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Chennai India

The increasing amount of underwater visual data and deep sea research have led to the rise of marine animal identification as a major field for data processing and analysis. The need to protect the ecosystem emphasizes how important this endeavor is, yet it’s still difficult because of things like the complexities of underwater backdrops, poor picture and video quality, and the variety of movements that marine life exhibits. This research used deep learning techniques to identify certain marine species and mark their presence with bounding boxes in uploaded images and videos in order to overcome these difficulties. The main goal is to provide workable methods for detecting marine life, so providing crucial backing for further studies in this area. By using the core deep learning frameworks, YOLOv8 and Faster-RCNN, the research was able to improve the efficiency and accuracy of marine life identification by making significant progress in detection time and mean average precision (mAP). This research work not only solves the current problems with finding marine animals but also establishes the groundwork for future research into comprehending and protecting the fragile undersea ecology.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Evaluating uniXcoder Embeddings for Automated Grading: A Study Across Varied Code Perspectives

Evaluating uniXcoder Embeddings for Automated Grading: A Stu...

引用

International Conference on computing and Networking Technology (ICCNT)

作者： Nakka Narmada Peeta Basa Pati Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Bengaluru India

ISBN: (数字)9798350370249

ISBN: (纸本)9798350370270

Automated grading systems play a pivotal role in modern educational settings, offering efficiency and scalability in assessing student submissions. In this study, we investigate the potential of uniXcoder embeddings of C language codes. A comparative analysis across four scenarios: code-only, codequestion, code-question-comment, and code-question-comment-reference is conducted. The dataset deals with C programs that are evaluated by Subject Matter Experts (SMEs). This study explores the analysis of code embeddings using K-Means approach and to determine the influence of Principal Component Analysis (PCA) and scaling techniques. The study compares clustering outcomes with and without PCA, with and without scaling techniques looking at measures including silhouette score, Davies-Bouldin index, Kullback-Leibler divergence and visualization techniques to draw conclusions. The findings suggest that uniXcoder embeddings demonstrate that “code-question-comment-reference” scenario is best suited for automated grading tasks. It also looks into the ability of the embedding to capture the level of error present in the code. This research contributes to the advancement of automated grading systems by providing insights into the suitability of uniXcoder embeddings for assessing student code submissions in educational environments.

关键词： Visualization Codes Subject matter experts Scalability C languages Indexes Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Graph-Based Approaches to Detect Network Anomalies and to Predict Attack Spread

Graph-Based Approaches to Detect Network Anomalies and to Pr...

引用

International Conference on computing and Networking Technology (ICCNT)

作者： S Rithuh Subhakkrith Y. Jagadish Kumar R Harish B Vishnu Dheeraj S Vidhya Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Coimbatore India

ISBN: (数字)9798350370249

ISBN: (纸本)9798350370270

Real world applications of data structures is enormous. Queues, stacks, trees, and graphs are commonly applied and implemented across various fields. To improve the efficiency of such implementations, optimisation through the underlying data structure can be considered. computer networks inherently are based on graphs with communicating systems as nodes and communication channels as edges. Information flows between the communicating systems through the communication channel. To detect suspicious connections and information flow, an optimal graph structure is implemented by modelling the information flow as edges, and retaining the communicating systems as nodes. Graph analytics is performed to capture graphical relationships, and the final graph structure is formed with the features of the information flow split into node features and edge features. Then this graph structure is passed to Graph Neural Networks (GNN) that captures hidden relationships of the graph, and transforms these graph features in such a way that basic Machine Learning models like k-Nearest Neighbors (KNN), Support Vector Classifier (SVC), and Random Forest can train themselves and model into better anomalous connection classifiers. Eleven baseline models are trained using the traditional way, and using GNN transformed embeddings, and their performances are observed. Testing shows that all models show an improvement in the evaluatory scores. Finally, a logic is provided to build an attack sequence graph to track the route in which the attack spreads among infected systems.

关键词： Image edge detection Communication channels Transforms Nearest neighbor methods Feature extraction Data structures Graph neural networks Vectors Security Random forests

来源：评论

学校读者我要写书评

暂无评论

Reliable and Fair Trustworthiness Evaluation Protocol for Platoon Service Recommendation System

引用

Chinese Journal of Electronics 2025年第2期34卷 563-575页

作者： Cheng Hongyuan Liu Yining Zhou Fei Tan Zhiyuan Zhang Xianchao School of Computer Science & Engineering Linyi University Linyi 273300 China School of Data Science and Artificial Intelligence Wenzhou University of Electronic Technology Wenzhou 325027 China Guangxi Key Laboratory of Digital Infrastructure Nanning 530023 China School of Computing Engineering & the Built Environment Edinburgh Napier University Edinburgh EH10 5DT UK College of Information Science and Engineering Key Laboratory of Medical Electronics and Digital Health of Zhejiang Province Engineering Research Center of Intelligent Human Health Situation Awareness of Zhejiang Province Jiaxing University Jiaxing 314001 China

Aiming at the problems of the communication inefficiency and high energy consumption in vehicular networks, the platoon service recommendation systems (PSRS) are presented. Many schemes for evaluating the reputation of platoon head vehicles have been proposed to obtain and recommend reliable platoon head vehicles. These trustworthiness evaluation protocols for PSRS fail to achieve both reliability and fairness. We first provide a reliable trustworthiness evaluation method to ensure that the reputation level of platoon head vehicle can be calculated by cloud service provider (CSP) with the help of key agreement mechanism and truth discovery technology. The semi-trusted entity CSP may maliciously tamper with the reputation level of the platoon head vehicle. We also provide a reputation level confirmation method to ensure the fairness of trustworthiness evaluation. Formal security proof and security analysis are provided to show that our trustworthiness evaluation protocol can achieve the goals of privacy protection, reliability, fairness and resistance to several security attacks. Experiments demonstrate that this protocol can save execution time and achieve reliable and fair trustworthiness evaluation for PSRS.

关键词： Reliability Fairness Privacy protection Trustworthiness evaluation Vehicle platoon

来源：评论

学校读者我要写书评

暂无评论

Automated Human Activity Recognition Using Attention Integrated Hybrid Conv-LSTM Model

Automated Human Activity Recognition Using Attention Integra...

引用

IEEE UP Section Conference on Electrical computer and Electronics (UPCON)

作者： Aira Udaybhasker Popuri Varun Kumar Tangaturu Devendranath Reddy Rimjhim Padam Singh Department of Computer Science & Engineering Amrita School of Computing Bengaluru Amrita Vishwa Vidyapeetham India

ISBN: (数字)9798350378726

ISBN: (纸本)9798350378733

Automated human activity recognition in videos has been of immense importance lately due to usage in the field of security and surveillance, patient monitoring, robotics, etc. Hence, this paper presents a novel hybrid model that combines the DenseNe201 convolutional neural network with Long Short Term Memory (LSTM) model and enhances it by incorporating multi-head attention mechanism for efficient Human Activity Recognition. The spatial feature is extracted using DenseNet201 architecture while the temporal relationship between the video frames is captured using LSTM layers and the features are assigned relative importance using the multi-head attention layers. The work analyzes the proposed model using standard UCF50 dataset and also trains other state-of-art models namely, Xception, MobileNetV2, InceptionV3, as well as a 3D CNN, etc with LSTM network on the dataset for comprehensive comparisons. The proposed hybrid approach has been found to deliver the best results as demonstrated by an F1-score of 80%. It also derives real-time activities from a single video input, essentially making the model practical.

关键词： Deep learning Analytical models Solid modeling Patient monitoring Surveillance Feature extraction Human activity recognition Convolutional neural networks Long short term memory Videos

来源：评论

学校读者我要写书评

暂无评论

MBTI Based Personality Prediction using Social Media Data

MBTI Based Personality Prediction using Social Media Data

引用

International Conference on computing and Networking Technology (ICCNT)

作者： R Prasanna Kumar G Bharathi Mohan M Rithani Gundala Pallavi Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Chennai India

ISBN: (数字)9798350370249

ISBN: (纸本)9798350370270

Existing methods for assessing an individual’s personality are very cumbersome due to those tests being very monotonous and erroneous due to a self-report bias. Whereas data-driven, automatic personality deduction raises privacy concern due to them needing external personal information. In our model, a person’s personality can be classified based on the 16 categories of Myers-Briggs personality types. With the availability of a huge amount of data on human behavior and personality traits, a machine learning model can be trained, and it can predict the personality trait of a person. A MBTI dataset is used to analyze and train the model. Upon analyzing the dataset, we found out that the classes are not balanced. To overcome this problem, we used binary classification. Next, we used the NLTK library to pre-process the data. We selected four machine learning models and trained our data in those models. Furthermore, we compared the results of four different machine learning models and chose the most accurate model. This research can be used to personalize advertisements, coaching for students and companies can use this research to attract specific types of audience.

关键词： Deep learning Training Logistic regression Accuracy Social networking (online) Companies Predictive models Data models Libraries Testing

来源：评论

学校读者我要写书评

暂无评论

A Novel Residual-Concatenate Feature Fusion Strategy for Building Segmentation From Remote Sensing Images

A Novel Residual-Concatenate Feature Fusion Strategy for Bui...

引用

Geoscience and Remote Sensing Symposium (M2GARSS), Mediterranean and Middle-East

作者： Akshara Preethy Byju Lakshmi Vijayan Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Amritapuri India

ISBN: (数字)9798350358582

ISBN: (纸本)9798350358599

This paper presents a novel encoder-decoder architecture based on U-Net integrated with a residual-concatenate multi-scale feature fusion (RCMFF) strategy to perform efficient building segmentation from remote sensing images. The proposed method considers an end-to-end deep learning framework based on UNet that consists of mainly three modules: i) an encoder with convolutional block attention module (CBAM) employed in their lateral connections; ii) the proposed residual-concatenate multi-scale feature fusion (RCMFF) module; and finally, iii) a decoder with refinement layers. The proposed RCMFF module integrates semantically relevant discriminative features inherent within the last encoder layer with reduced number of parameters thereby simultaneously improving the segmentation accuracy as well as the computational speed. The effectiveness of the proposed module is assessed using two benchmark archives: Massachusetts (MAS) and WHU building datasets. The proposed model demonstrates an overall improvement of 2.30% and 7.01% compared to the baseline models DeepLabV3+ and UNet++, respectively.

关键词： Deep learning Image segmentation Shape Computational modeling Architecture Buildings Urban areas

来源：评论

学校读者我要写书评

暂无评论

Automated Skin Cancer Detection Using Convolutional Neural Networks: A Deep Learning Approach for Dermoscopic Image Analysis

Automated Skin Cancer Detection Using Convolutional Neural N...

引用

BIT Conference (BITCON), International

作者： Rekha R Nair Ebin P M Tina Babu Department of Computer Science and Engineering Alliance School of Advanced Computing Alliance University Bengaluru India

ISBN: (数字)9798331518394

ISBN: (纸本)9798331518400

Skin cancer is a serious global health concern, highlighting the necessity of early detection and treatment. A specialized Convolutional Neural Network (CNN) model was created to automate the diagnosis of skin cancer in dermoscopic pictures. A diverse collection of skin lesion photos was used to train the model using a pre-trained ResNet50 architecture with custom layers. It scored a test accuracy of 91.5%, precision of 89.3%, recall of 93.2%, and an F1-score of 91.2%. The model’s AUC-ROC value of 0.957 demonstrates its exceptional ability to discriminate between benign and malignant tumors. Comparative analysis showed the proposed CNN outperformed traditional machine learning methods and achieved comparable results to other deep learning approaches. Hyperparameter tuning further improved model performance, with the best configuration achieving 93.2% accuracy. The study also explored techniques for model interpretability to enhance clinical applicability. The findings highlight the effectiveness of CNN-based models in achieving accurate and efficient skin cancer detection. This study supports the advancement of automated diagnostic tools, which can aid in early detection and contribute to better patient outcomes.

关键词： Deep learning Support vector machines Accuracy Malignant tumors Skin Convolutional neural networks Lesions Tuning Skin cancer Residual neural networks

来源：评论

学校读者我要写书评

暂无评论

A Novel Approach for Phishing Website Detection using Deep Learning Models

A Novel Approach for Phishing Website Detection using Deep L...

引用

International Conference on Computation System and Information Technology for Sustainable Solutions (CSITSS)

作者： Chellu Vyshnavi Chukka Mishal Raj Gokul Gopakumar Guddeti Bindu Prasanna Suja Palaniswamy Department of Computer Science & Engineering Amrita School of Computing Bengaluru Amrita Vishwa Vidyapeetham India

ISBN: (数字)9798331505462

ISBN: (纸本)9798331505479

Phishing website detection is a major task in cybersecurity, aimed at saving users from sites that steal personal information. The traditional approaches can hardly minimize false positives and cope with the dynamic strategies of phishing. In this work, a novel approach has been proposed to detect phishing websites using deep learning models, such as CNN, RNN, RNN with attention layers, CRNN, LSTM, and GRU, to enhance accuracy in detection. This work has addressed one vital domain challenge, which includes very rapid changes of techniques and phishing activities, and the real-time detection feature that conventional models generally may not have. The proposed approach will leverage these advanced models for automatic feature extraction and adaptability to new strategies for significant improvement of detection accuracy and reduction of false positives. The proposed method gives a robust and real-time solution for the task of phishing website detection and improving cybersecurity in general, after its evaluation on a comprehensive dataset.

关键词： Deep learning Adaptation models Accuracy Phishing Computational modeling Feature extraction Real-time systems Information technology Long short term memory

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：