检索结果-内蒙古大学图书馆

Optimizing deep learning architectures for novel view synthesis: Investigating the impact of NeRF MLP parameters on complex scenes

引用

International Journal of Information Technology (Singapore) 2024年第4期16卷 2295-2305页

作者： Singla, Kanika Nand, Parma Department of Computer Science & Engineering School of Engineering and Technology Sharda University Greater Noida India

Machine learning combined with geometric reasoning is a promising approach for generating new perspectives of a scene using limited image captures, known as neural rendering techniques. Neural radiance fields (NeRF) represent complex 3D scenes by learning the radiance implicit function and have shown impressive performance in various applications. However, the influence of NeRF's parameters on its overall performance remains largely unexplored. In this paper, we conduct an extensive grid search on NeRF's multi-layer perceptron (MLP) parameters to examine their impact on image quality. We analyze the results by varying parameters related to the MLP architecture, such as hidden layer count and neurons per layer. Our findings reveal that the choice of MLP architecture significantly affects image quality. Our proposed approach outperforms NeRF's default settings on challenging scenes, including synthetic and real-world datasets. These results shed light on the impact of different parameters on NeRF's performance, assisting practitioners in selecting suitable settings for their specific applications. © 2023, The Author(s), under exclusive licence to Bharati Vidyapeeth's Institute of computer Applications and Management.

关键词： Neural radiance fields (NeRF) Neural rendering Novel view synthesis Peak Signal to Noise Ratio (PSNR)

来源：评论

学校读者我要写书评

暂无评论

Eye-Tracking Based Autism Spectrum Disorder Diagnosis Using Chaotic Butterfly Optimization with Deep Learning Model

引用

computers, Materials & Continua 2023年第8期76卷 1995-2013页

作者： Tamilvizhi Thanarajan Youseef Alotaibi Surendran Rajendran Krishnaraj Nagappan Department of Computer Science and Engineering Panimalar Engineering CollegeChennai600123India Department of Computer Science University College of Al JamoumUmm Al-Qura UniversityMakkah21421Saudi Arabia Department of Computer Science and Engineering Saveetha School of EngineeringSaveetha Institute of Medical and Technical SciencesChennai602105India School of Computing SRM Institute of Science and TechnologyKattankulathur603203India

Autism spectrum disorder(ASD)can be defined as a neurodevelopmental condition or illness that can disturb kids who have heterogeneous characteristics,like changes in behavior,social disabilities,and difficulty communicating with *** tracking(ET)has become a useful method to detect *** vital aspect of moral erudition is the aptitude to have common visual *** eye-tracking approach offers valuable data regarding the visual behavior of children for accurate and early ***-tracking data can offer insightful information about the behavior and thought processes of people with ASD,but it is important to be aware of its limitations and to combine it with other types of data and assessment techniques to increase the precision of ASD *** operates by scanning the paths of eyes for extracting a series of eye projection points on images for examining the behavior of children with *** purpose of this research is to use deep learning to identify autistic disorders based on eye *** Chaotic Butterfly Optimization technique is used to identify this specific ***,this study develops an ET-based Autism Spectrum Disorder Diagnosis using Chaotic Butterfly Optimization with Deep Learning(ETASD-CBODL)*** presented ETASDCBODL technique mainly focuses on the recognition of ASD via the ET and DL *** accomplish this,the ETASD-CBODL technique exploits the U-Net segmentation technique to recognize interested *** addition,the ETASD-CBODL technique employs Inception v3 feature extraction with CBO algorithm-based hyperparameter ***,the long-shorttermmemory(LSTM)model is exploited for the recognition and classification of *** assess the performance of the ETASD-CBODL technique,a series of simulations were performed on datasets from the figure-shared data *** experimental values of accuracy(99.29%),precision(98.78%),sensitivity(99.29%)and specificity(99.29%)showed a better perfo

关键词： Eye tracking ASD diagnosis deep learning butterfly optimization algorithm behavior analysis

来源：评论

学校读者我要写书评

暂无评论

Task offloading delay minimization in vehicular edge computing based on vehicle trajectory prediction

引用

Digital Communications and Networks 2025年第2期11卷 537-546页

作者： Feng Zeng Zheng Zhang Jinsong Wu School of Computer Science and Engineering Central South UniversityChangsha410083China School of Artificial Intelligence Guilin University of Electronic TechnologyGuilin541000China The Department of Computer Science Universidad de ChileSantiagoChile

In task offloading,the movement of vehicles causes the switching of connected RSUs and servers,which may lead to task offloading failure or high service *** this paper,we analyze the impact of vehicle movements on task offloading and reveal that data preparation time for task execution can be minimized via forward-looking ***,a Bi-LSTM-based model is proposed to predict the trajectories of *** service area is divided into several equal-sized *** the actual position of the vehicle and the predicted position by the model belong to the same grid,the prediction is considered correct,thereby reducing the difficulty of vehicle trajectory ***,we propose a scheduling strategy for delay optimization based on the vehicle trajectory *** the inevitable prediction error,we take some edge servers around the predicted area as candidate execution servers and the data required for task execution are backed up to these candidate servers,thereby reducing the impact of prediction deviations on task offloading and converting the modest increase of resource overheads into delay reduction in task *** results show that,compared with other classical schemes,the proposed strategy has lower average task offloading delays.

关键词： Vehicular edge computing Task offloading Vehicle trajectory prediction Delay minimization Bi-LSTM model

来源：评论

学校读者我要写书评

暂无评论

Semi-supervised Domain Adaptation for Semantic Segmentation via Active Learning with Feature- and Semantic-Level Alignments

IEEE Transactions on Intelligent Vehicles

引用

IEEE Transactions on Intelligent Vehicles 2024年 1-11页

作者： Wen, Lu Xu, Yuanyuan Feng, Zhenghao Zhou, Jiliu Zhou, Luping Wang, Yan School of Computer Science Sichuan University China School of Electrical and Information Engineering The University of Sydney Sydney NSW Australia

Unsupervised domain adaptation (UDA) is a popular technique to reduce the manual annotation cost in semantic segmentation. However, due to the absence of strong supervision in the target domain, UDA is prone to biasing the decision boundary towards the source domain. To alleviate this issue, this paper proposes a more effective semi-supervised domain adaptation (SSDA) method for semantic segmentation via active learning with feature- and semantic-level alignments. Specifically, active learning is utilized to select those samples with high diversity and uncertainty from the target domain for labeling. These selected data could provide reliable clues for domain transfer since they reveal the intrinsic distribution of the target domain as well as including hard samples at boundaries. Moreover, to better adapt the segmentation model from the source data to the labeled target data selected above, we propose a scheme based on both feature- and semantic-level domain alignments. The feature-level domain alignment imposes the distribution consistency between the Transformer features of the two domains by adversarial learning, which is a global alignment. In contrast, the semantic-level domain alignment optimizes the affinity and divergence of the semantic representations across domains via contrastive learning, which is a local alignment. These two alignments jointly bridge the domain gap from both the global and the local views, respectively. In addition, the pseudo labels of the unlabeled data are generated to expand the labeled data and further strengthen the cross-domain segmentation in a self-training manner. Extensive experiments on segmentation benchmarks demonstrate the effectiveness of our proposed method. IEEE

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Intelligent waste classification approach based on improved multi-layered convolutional neural network

引用

Multimedia Tools and Applications 2024年第36期83卷 84095-84120页

作者： Chhabra, Megha Sharan, Bhagwati Elbarachi, May Kumar, Manoj Department of Computer Science & amp Engineering School of Engineering & ampTechnology Sharda University UP Gr. Noida India School of Engineering and Sciences Department of Computer Science and Engineering SRM University-AP Andhra Pradesh Amaravati India Head of School of Computer Science University of Wollongong in Dubai Dubai Knowledge Park Dubai United Arab Emirates School of Computer Science University of Wollongong in Dubai Dubai Knowledge Park Dubai United Arab Emirates MEU Research Unit Middle East University Amman11831 Jordan

This study aims to improve the performance of organic to recyclable waste through deep learning techniques. Negative impacts on environmental and Social development have been observed relating to the poor waste segregation schemes. Separating organic waste from recyclable waste can lead to a faster and more effective recycling process. Manual waste classification is a time-consuming, costly, and less accurate recycling process. Automated segregation in the proposed work uses Improved Deep Convolutional Neural Network (DCNN). The dataset of 2 class category with 25077 images is divided into 70% training and 30% testing images. The performance metrics used are classification Accuracy, Missed Detection Rate (MDR), and False Detection Rate (FDR). The results of Improved DCNN are compared with VGG16, VGG19, MobileNetV2, DenseNet121, and EfficientNetB0 after transfer learning. Experimental results show that the image classification accuracy of the proposed model reaches 93.28%. © The Author(s) 2024.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Transfer Learning Based Yogic Posture Recognition System Using Deep Pre-trained Features

引用

SN computer science 2024年第6期5卷 1-13页

作者： Rajendran, Arun Kumar Sethuraman, Sibi Chakkaravarthy School of Computer Science and Engineering VIT-AP University School of Computer Science & amp Engineering and Center of Excellence Artificial Intelligence and Robotics (AIR) VIT-AP University

Yoga is a healthy exercise that focuses on physical, psychological, and divine connections. However, engaging in yoga while adopting poor postures might result in health issues like muscle discomfort and sprains. Discovering the specific posture amid an activity is still a very challenging and complicated task in artificial intelligence. With the aid of transfer learning, a model which can identify the yogic posture from images has been constructed in this work. In this approach, sixteen postures of yogic data are collected, and then pre-processing techniques are used to resize the images. This work employs the transfer learning method to extract the features from popular convolutional neural network architectures such as VGG16, VGG19, InceptionV3, Xception, MobileNet, and MobileNetV2;all are pre-trained on the ImageNet database. Extracted features are fed to machine learning models like Naive Bayes (NB), Support Vector Machine (SVM), Random Forest (RF), and Extreme Gradient Boosting (XGBoost) for classification. The linear SVM-based classifier, trained on features extracted by MobileNet, performs better than its rivals with an accuracy of 99.92%. In total, six transfer learning models are exploited in this yogic posture classification approach and yield encouraging results in terms of prediction accuracy. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： InceptionV3 MobileNet MobileNetV2 NB RF SVM Transfer learning VGG16 VGG19 Xception XGBoost Yoga posture classification

来源：评论

学校读者我要写书评

暂无评论

Rank-Based singular value decomposition recommendation algorithm for academic web pages

引用

International Journal of computers and Applications 2025年第5期47卷 413-423页

作者： Naseer, Mehwish Zhang, Wu National University of Sciences and Technology Islamabad Pakistan School of Computer Engineering and Science Shanghai University Shanghai China

Finding appropriate information on the web is a tedious task and thus demands an intelligent mechanism to assist users for this purpose. Students are the victims of information overloading on the internet the most, as it hinders their learning. This study proposes a rank-based singular value decomposition (R-SVD) algorithm for making academic web page recommendations. The weblog of the network is a depiction of student's browsing habits. The proposed approach takes weblog data as input. The rank of the pages is calculated based on the hit ratio of each page. Rank values are normalized to get a scaled value of each page rank. Then R-SVD is used with multiple similarities and recommendation results are evaluated based on the different parameters such as by computation of root mean square error (RMSE), mean absolute error (MAE) and mean percent error (MPE). The proposed approach recommends web pages with the least error values and verdict that the proposed method is a justifiable technique for the academic web page recommendation system. The proposed approach is compared with multiple recommendation algorithms such as User-KNN, Item-KNN, Non-Negative Matrix Factorization (NMF), Bayesian User Community Model (BUCM), and Probabilistic Matrix Factorization (PMF) using numerous similarity measures. © 2025 Informa UK Limited, trading as Taylor & Francis Group.

关键词： Non-negative matrix factorization

来源：评论

学校读者我要写书评

暂无评论

Parameter Efficient Point Cloud Prompt Tuning for Unified Point Cloud Understanding

IEEE Transactions on Intelligent Vehicles

引用

IEEE Transactions on Intelligent Vehicles 2024年 1-16页

作者： Fei, Ben Liu, Liwen Yang, Weidong Li, Zhijun Chen, Wen-Ming Ma, Lipeng School of Computer Science Fudan University Shanghai China School of Mechanical Engineering Tongji University Shanghai China Academy for Engineering and Technology Fudan University Shanghai China

The prevailing paradigm in 3D vision involves fully fine-tuning all the backbone parameters of pre-trained models. However, this approach poses challenges due to the large number of parameters requiring tuning, resulting in unexpected storage demands. To address these issues and alleviate the computational cost and storage burden associated with full fine-tuning, we propose Point Cloud Prompt Tuning (PCPT) as an effective method for large Transformer models in point cloud processing. PCPT offers a powerful and efficient solution to mitigate the costs associated with full fine-tuning. Drawing inspiration from recent advancements in efficient tuning of large-scale language models and 2D vision models, PCPT leverages less than 0.05 % of trainable parameters, while keeping the pre-trained parameters of the Transformer backbone unchanged. To evaluate the effectiveness of PCPT, extensive experiments were conducted on four discriminative datasets (ModelNet40, few-shot ModelNet40, ScanObjectNN, ShapeNetPart) and four generation datasets (PCN, MVP, ShapeNet55, and ShapeNet34/Unseen21). The results demonstrate that the task-specific prompts utilized in PCPT enable the Transformer model to adapt effectively to the target domains, yielding results comparable to those obtained through other full fine-tuning methods. This highlights the versatility of PCPT across various domains and tasks. Our code is available at https://***/Fayeben/PCPT. IEEE

关键词： Three dimensional displays

来源：评论

学校读者我要写书评

暂无评论

ArchSentry: Enhanced Android Malware Detection via Hierarchical Semantic Extraction

引用

IEEE Transactions on Network and Service Management 2025年第3期22卷 2822-2837页

作者： Wang, Tianbo Liu, Mengyao Li, Huacheng Zhao, Lei Jiang, Changnan Xia, Chunhe Cui, Baojiang Beihang University School of Cyber Science and Technology Beijing100191 China Beihang University School of Computer Science and Engineering Beijing China Wuhan University School of Cyber Science and Engineering Wuhan China Beijing University of Posts and Telecommunications School of Cyberspace Security Beijing China

Android malware poses a significant challenge for mobile platforms. To evade detection, contemporary malware variants use API substitution or obfuscation techniques to hide malicious activities and mask their shallow semantic characteristics. However, existing research lacks analysis of the hierarchical semantic associated with Android apps. To address this problem, we propose ArchSentry, an enhanced Android malware detection via hierarchical semantic extraction. First, we select entities and their relationships relevant to Android software behavior through the software architecture and represent them using a heterogeneous graph. Then, we structure meta-paths to represent rich semantic information to achieve semantic enhancement and improve efficiency. Next, we design a meta-path semantic selection method based on KL Divergence to identify and eliminate redundant features. To achieve a comprehensive representation of the overall software semantics and improve performance, we construct a feature fusion approach based on Restricted Boltzmann Machines (RBM) and AutoEncoder (AE) during the pre-training phase, while preserving the probability distribution characteristics of various meta-paths. Finally, Deep Neural Networks (DNN) process fusion features for comprehensive feature sets. Experimental results on real-world application samples indicate that ArchSentry achieves a remarkable 99.2% detection rate for Android malware, with a low false positive rate below 1%. These results surpass the performance of current state-of-the-art approaches. © 2004-2012 IEEE.

关键词： Android malware

来源：评论

学校读者我要写书评

暂无评论

Feature Selection and Support Vector Machine Classification method for Banknote Dirtiness Recognition Based on Marine Predator Algorithm with Mathematical Functions

引用

Journal of Intelligent and Fuzzy Systems 2023年第3期45卷 4315-4336页

作者： Guo, Fu-Jun Sun, Wei-Zhong Wang, Jie-Sheng Zhang, Min Hou, Jia-Ning Song, Hao-Ming Wang, Yu-Cai School of Electronic and Information Engineering University of Science and Technology Liaoning Anshan China School of Computer Science and Software Engineering University of Science and Technology Liaoning Anshan China

Dealing with classification problems requires the crucial step of feature selection (FS), which helps to reduce data dimensions and shorten classification time. Feature selection and support vector machines (SVM) classification method for banknote dirtiness recognition based on marine predator algorithm (MPA) with mathematical functions was proposed. The mathematical functions were mainly used to improve the optimizatio of MPA for feature parameter selection, and the loss function and kernel function parameters of the SVM are optimized by slime mold optimization algorithm (SMA) and marine predator algorithm. According to the experimental results, the accuracy of identifying dirtiness on the entire surface of the banknote reaches 89.07%. At the same time, according to the image pattern distribution of the banknoteS, the white area image in the middle left of the collected banknote is selected by the same method to select the feature parameters and identify the dirtiness of the banknoteS. The accuracy of dirtiness recognition in the middle left white area reached 86.67%, this shows that the white area in the middle left can basically completely replace the entire banknote. To confirm the effectiveness of the feature selection method, the proposed optimization method has been compared with four other swarm intelligent optimization algorithms to verify its performance. The experiment results indicate that the enhanced strategy is successful in improving the performance of MPA. Moreover, the robustness analysis proves its effectiveness. © 2023 - IOS Press. All rights reserved.

关键词： Support vector machines

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：