检索结果-内蒙古大学图书馆

2nd IEEE International Conference on Networking and Communications, ICNWC 2024

作者： Modak, Masooda Temgire, Devdatta Kaushik, Serin Kachare, Vaibhav Shriyan, Dheeraj SIES Graduate School of Technology Department of Computer Engineering Nerul India

ISBN: (纸本)9798350365269

In today's fast-paced educational landscape, understanding and addressing students' stress levels are paramount for fostering effective learning environments. This research introduces a novel approach to stress detection and adaptive reading, leveraging the Internet of Things (IoT), galvanic skin response (GSR) sensors, and machine learning techniques. By collecting real-time physiological data from students wearing GSR sensors during tests, our system aims to dynamically adapt reading materials based on individual stress levels. This paper details the methodology, including the creation of a customized dataset, data preprocessing, clustering, and the development of an IoT infrastructure using Raspberry Pi. The integration of Django web applications, WebSocket communication, and machine learning models further enhances the adaptability and responsiveness of the system. Our work contributes to the intersection of IoT, GSR, and machine learning in education, offering a promising avenue for personalized and stress-aware learning experiences. © 2024 IEEE.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Multi-Branch Fusion Module for Siamese Object Tracking 2

Multi-Branch Fusion Module for Siamese Object Tracking

引用

2nd International Conference on Informatics, Networking, and Computing, ICINC 2023

作者： Li, Gaoxiang Lu, Xiaofeng Wang, Chongke Xu, Jingbo School of Computer Science and Engineering Xi`an University of Technology Xi’an China

ISBN: (纸本)9781510674745

Various mainstream target tracking algorithms based on siamese networks are gradually becoming a trend in the field of deep learning tracking due to their concurrent advantages of accuracy and speed. Most siamese network-based trackers describe the tracking of a target object as a similar matching problem, and these trackers have achieved more advanced performance in several public tests. Most trackers often suffer from tracking drift or performance degradation owing to the non-updating of the template in the first frame and the target appearance encounters disturbing environments such as occlusion and drastic deformation. To address the current problem, this paper proposes an algorithm for fusion of multi-branch modules based on siamese network to achieve the fusion of target templates for updating, which improves the tracker's anti-jamming ability and the network structure is anchor-free. Meanwhile, this paper designs a new fusion module that fuses templates by inserting weight tensor method for multi-template fusion and optimises the results by complementary weight tensors. This method is practiced in SiamFC++ algorithm, where the target dataset is input and features are extracted, and then classification and regression operations are performed by fusion of multi-branch modules to get the predicted position of the target. © 2024 SPIE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

A Novel Hybrid Approach for Brain Tumour MRI Segmentation Leveraging Transfer Learning 5

A Novel Hybrid Approach for Brain Tumour MRI Segmentation Le...

引用

5th IEEE Global Conference for Advancement in technology, GCAT 2024

作者： Sudheer Kumar, E. Anumandla, Sai Keerthana Krishnakumar, Anagha Bhavana, Gondi Vidya Vellore Institute of Technology School of Computer Science and Engineering Chennai India

ISBN: (纸本)9798350376685

Integrating multiple deep learning architectures often significantly improves the performance of a model. This approach is particularly beneficial for challenging tasks like brain tumour detection. This paper provides a detailed examination of a novel DenseNet-ResNet fusion model that leverages DenseNet's strength in feature extraction and ResNet's robust learning process. This helps in overcoming the limitations of each separate model and, hence, leads to increased accuracy and improved training efficiency. Additionally, various augmentation techniques were applied to the data to improve the performance;these included brightness adjustments, rotation, shifting, shear transformation, zooming, and horizontal flipping. We also employed histogram equalization, to improve contrast, making the image clearer for visualizing the tumours. Furthermore, image registration algorithms were embedded for the alignment of flipped images, and Grad-CAM visualization was utilized for a more comprehensive evaluation of the results. Our results have shown that the DenseNet-ResNet fusion model not only gives higher classification accuracy but is also better at generalization. © 2024 IEEE.

关键词： Image registration

来源：评论

学校读者我要写书评

暂无评论

A novel composite machine learning model for the prediction of compressive strength of blended concrete

引用

Journal of Building Pathology and Rehabilitation 2025年第1期10卷 1-18页

作者： Prasad, E.V. Rama Krishna, S. Singha, S. Department of Civil Engineering School of Technology GITAM (Deemed to be University) Telangana Hyderabad India Department of Computer Science and Engineering School of Technology GITAM (Deemed to be University) Telangana Hyderabad India

Blended cement, comprising clinker and supplementary cementitious materials (SCMs), such as fly ash, slag, and silica fume, forms blended concrete when combined with aggregates. This study introduced a novel ensemble approach, the LightGBM and ANN Fusion (LAF) model, to check the prediction of compressive strength in blended concrete using eight key input features, including cement content, slag, fly ash, water, and curing age, derived from 1030 samples. The LAF model synergistically combines predictions from LightGBM and ANN with the fusion weight (alpha) dynamically optimized using Support Vector Machine (SVM) analysis. This integration leverages the strengths of both models while mitigating their individual limitations. The LAF model exhibited superior performance, with an R² of 0.9514 during training and 0.8863 during testing, significantly surpassing standalone models such as LightGBM (R² = 0.9385) and ANN (R² = 0.9136). It achieved the lowest error rates, including an MSE of 15.7384 and RMSE of 3.9672 during training, underscoring its robust generalization capabilities. The novel utilization of SVM-tuned alpha for dynamic model adaptation reduces overfitting and enhances predictive accuracy compared to traditional ensemble and single models. Comparative analysis with existing ensemble techniques, including stacking and gradient-boosted ANN, further validated the superior accuracy and adaptability of the LAF model. This approach represents a significant advancement in predictive modeling, offering a scientifically sound and versatile framework for concrete strength prediction. Future research will refine the SVM tuning, expand the model’s application to other material properties, and explore real-time deployment, positioning the LAF model as a pioneering tool for smart construction materials. © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2024.

关键词： ANN Compressive strength LAF LGBM Performance indicators Sensitivity analysis SVM

来源：评论

学校读者我要写书评

暂无评论

Fusion of color and hallucinated depth features for enhanced multimodal deep learning-based damage segmentation

引用

Earthquake engineering and engineering Vibration 2023年第1期22卷 55-68页

作者： Tarutal Ghosh Mondal Mohammad Reza Jahanshahi Civil Architectural and Environmental EngineeringMissouri University of Science and TechnologyRollaMOUSA Lyles School of Civil Engineering Purdue UniversityWest LafayetteINUSA School of Electrical and Computer Engineering Purdue UniversityWest LafayetteINUSA

Recent advances in computer vision and deep learning have shown that the fusion of depth information can significantly enhance the performance of RGB-based damage detection and segmentation ***,alongside the advantages,depth-sensing also presents many practical *** instance,the depth sensors impose an additional payload burden on the robotic inspection platforms limiting the operation time and increasing the inspection ***,some lidar-based depth sensors have poor outdoor performance due to sunlight contamination during the *** this context,this study investigates the feasibility of abolishing depth-sensing at test time without compromising the segmentation *** autonomous damage segmentation framework is developed,based on recent advancements in vision-based multi-modal sensing such as modality hallucination(MH)and monocular depth estimation(MDE),which require depth data only during the model *** the time of deployment,depth data becomes expendable as it can be simulated from the corresponding RGB *** makes it possible to reap the benefits of depth fusion without any depth perception per *** study explored two different depth encoding techniques and three different fusion strategies in addition to a baseline RGB-based *** proposed approach is validated on computer-generated RGB-D data of reinforced concrete buildings subjected to seismic *** was observed that the surrogate techniques can increase the segmentation IoU by up to 20.1%with a negligible increase in the computation ***,this study is believed to make a positive contribution to enhancing the resilience of critical civil infrastructure.

关键词： multimodal data fusion depth sensing vision-based inspection UAV-assisted inspection damage segmentation post-disaster reconnaissance modality hallucination monocular depth estimation

来源：评论

学校读者我要写书评

暂无评论

Improved Transfer Learning based Deep Convolutional Neural Network for Lung Disease Classification using CT image 3

Improved Transfer Learning based Deep Convolutional Neural N...

引用

3rd International Conference for Advancement in technology, ICONAT 2024

作者： Shafi, Syed Mohammed Chinnappan, Sathiya Kumar Vellore Institute of Technology School of Computer Science and Engineering Vellore India

ISBN: (纸本)9798350354171

One of the most terrible diseases that affect many nations on the globe is lung-related diseases, and early diagnosis of the disease is still a difficult task. It takes time and extra human effort for the oncologists to evaluate the lung disease based on the findings of blood tests and CT scans. To lower mortality, an automated approach for accurate identification of lung diseases should be proposed. Even though several researchers have put out approaches for detecting lung diseases, the methods now in use for early-stage disease detection still do not fulfil the expectations. This research developed a new lung disease classification (LDC) approach in CT images using the Transfer learning concept. At first, the CT image is pre-processed using the wiener filter (WF) method. The classification of lung disease is performed using Transfer Learning based Improved Deep CNN (TL-IDCNN) framework. © 2024 IEEE.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

Metal Surface Defect Detection Using Deep Learning Techniques

Metal Surface Defect Detection Using Deep Learning Technique...

引用

2024 International Conference on Signal Processing, Computation, Electronics, Power and Telecommunication, IConSCEPT 2024

作者： Abhinav, S. Elakiya, E. School of Computer Science and Engineering Vellore Institute of Technology Chennai India

ISBN: (纸本)9798331540685

Metal defect detection plays a crucial role in ensuring the quality and safety of industrial manufacturing processes. Traditional methods of defect detection often rely on manual inspection, which can be time-consuming and prone to human error. In recent years, the application of deep learning techniques, particularly Convolutional Neural Networks (CNNs), has shown promising results in automating defect detection tasks. This research explores the effectiveness of using CNN architectures, specifically InceptionNet, ResNet34, and VGG16, for automated metal defect detection. The study involves training and evaluating these models on a dataset of metal surface images containing various types of defects. The performance metrics such as accuracy, precision, recall, and F1-score are analysed to assess the robustness and efficacy of each model in defect classification. Additionally, the research investigates the impact of transfer learning by fine-tuning pre-trained CNN models on the defect detection task. The experimental results demonstrate the superior performance of deep CNNs compared to traditional methods, highlighting the potential of these techniques for real-world applications in industrial quality control and inspection. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Low-Latency State Management for Real-Time Tasks in Edge Serverless 22

Low-Latency State Management for Real-Time Tasks in Edge Ser...

引用

22nd IEEE International Symposium on Parallel and Distributed Processing with Applications, ISPA 2024

作者： Wen, Yanyan Xu, Guangping Wang, Jianshe Hao, Wei Tianjin University of Technology School of Computer Science and Engineering Tianjin China

ISBN: (纸本)9798331509712

Stateful serverless systems commonly adopt an architectural paradigm characterized by compute and storage separation within cloud data centers. Nevertheless, guaranteeing prompt response for real-time tasks at the edge becomes challenging due to network overheads. This paper introduces a Low-Latency state management framework for real-time tasks in edge serverless systems called LoLa, which adaptively places states proximate to functions, thereby mitigating delays in accessing states within edge serverless systems. Our approach aims at mitigating network latency and optimizing resource utilization by co-locating functions and states, thereby enhancing the system's overall efficiency. We introduce an adaptive strategy to coordinate the migration of states. It dynamically adjusts the positions of states based on historical data and real-time feedback. Additionally, we designed an in-memory state storage mechanism to facilitate low-latency access and implement a lightweight and fine-grained state management to ensure stored state consistency. Evaluation results showcase the efficacy of LoLa in reducing state read and write latency within edge serverless systems. Specifically, the average response latency is observed to decrease by 65.2% and 38.1% in the best and worst-case scenarios, respectively. © 2024 IEEE.

关键词： Edge computing Stateful Serverless

来源：评论

学校读者我要写书评

暂无评论

PSR-Tree: A Novel Method for Personalized Trajectory Data Protection 22

PSR-Tree: A Novel Method for Personalized Trajectory Data Pr...

引用

22nd IEEE International Symposium on Parallel and Distributed Processing with Applications, ISPA 2024

作者： Wang, Chundong Song, Yixuan Zhao, Yongxin Tian, Yuhan Fang, Shunyao Tianjin University of Technology School of Computer Science and Engineering Tianjin China

ISBN: (纸本)9798331509712

Privacy concerns related to location-based services are becoming increasingly critical. To address the limitations of previous research, which has often neglected the temporal attributes of trajectories and relied solely on average noise addition methods, a trajectory data publishing scheme is proposed based on an R-tree storage structure. Initially, trajectory data is dimensionalized using a spatiotemporal optimal trajectory segmentation algorithm, and the preprocessed key trajectory data is organized into a PSR-Tree. Subsequently, an incremental privacy budget allocation strategy is applied to add noise to the trajectory data. Finally, a Hidden Markov Model is employed to constrain noise addition at the PSR-Tree nodes, fulfilling the requirements for trajectory data publishing. Experimental comparisons with existing methods demonstrate that this approach effectively addresses the issue of insufficient data usability and exhibits robust scalability. © 2024 IEEE.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

Multi-Scale Prompt Memory-Augmented Model for Black-Box Scenarios

Multi-Scale Prompt Memory-Augmented Model for Black-Box Scen...

引用

2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL 2024

作者： Kuang, Xiaojun Philip Chen, C.L. Li, Shuzhen Zhang, Tong School of Computer Science & Engineering South China University of Technology China

ISBN: (纸本)9798891761148

Black-box few-shot text classification handles text classification in limited data without accessing the parameters and gradients of language models (LMs). Existing black-box optimization methods have demonstrated strong few-shot learning capabilities. However, they still require numerous LMs’ calls to search optimal prompts, thus resulting in overfitting performance and increasing computational cost. To address this issue, we present MuSKPrompt (Multi-scale Knowledge Prompt for Memory Model), an efficient multi-scale knowledge prompt-based memory model in black-box few-shot text classification task. MuSKPrompt extracts instance-level and class-level knowledge at different scales and stores them in memory banks during training. Then, it references multi-scale memory banks to perform quick inference on new samples via a novel scoring module. MuSKPrompt achieves competitive performance in limited data through multi-scale instance-level and class-level knowledge. Moreover, it realizes gradient-free optimization with zero training parameters in the black-box scenario. Experiments on different benchmarks and parameter analysis demonstrate the effectiveness and efficiency of MuSKPrompt in black-box few-shot text classification tasks. ©2024 Association for Computational Linguistics.

关键词： Classification (of information)

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：