检索结果-内蒙古大学图书馆

artificial intelligence-assisted visual inspection for cultural heritage: State-of-the-art review

JOURNAL OF CULTURAL HERITAGE 2024年 66卷 536-550页

作者： Mishra, Mayank Lourenco, Paulo B. Univ Minho Dept Civil Engn ISISE ARISE P-4804533 Guimaraes Portugal

Applying computer science techniques such as artificial intelligence (AI), deep learning (DL), and computer vision (CV) on digital image data can help monitor and preserve cultural heritage (CH) sites. Defects such as weathering, removal of mortar, joint damage, discoloration, erosion, surface cracks, vegetation, seepage, and vandalism and their propagation with time adversely affect the structural health of CH sites. Several studies have reported damage detection in concrete and bridge structures using AI techniques. However, few studies have quantified defects in CH structures using the AI paradigm, and limited case studies exist for their applications. Hence, the application of AI-assisted visual inspections for CH sites needs to be explored. AI-assisted digital inspections assist inspection professionals and increase confidence levels in the damage assessment of CH buildings. This review summarizes the damage assessment techniques using image processing techniques, focusing mainly on DL techniques applied for CH conservation. Several case study applications of CH buildings are presented where AI can assist in traditional visual inspections. (c) 2024 The Author(s). Published by Elsevier Masson SAS on behalf of Consiglio Nazionale delle Ricerche (CNR). This is an open access article under the CC BY license ( http://***/licenses/by/4.0/ )

关键词： Deep learning UAV Damage identification Convolutional neural networks Computer vision Structural health monitoring

来源：评论

学校读者我要写书评

暂无评论

PETNet- Coincident Particle Event Detection using Spiking neural networks

PETNet- Coincident Particle Event Detection using Spiking Ne...

引用

Neuro Inspired Computational Elements Conference (NICE)

作者： Debus, Jan Debus, Charlotte Dissertori, Guenther Goetz, Markus Swiss Fed Inst Technol Inst Particle Phys & Astrophys Zurich Switzerland Karlsruhe Inst Technol Sci Comp Ctr Karlsruhe Germany Helmholtz AI Munich Germany

ISBN: (纸本)9798350390599;9798350390582

Spiking neural networks (SNN) hold the promise of being a more biologically plausible, low-energy alternative to conventional artificial neural networks. Their time-variant nature makes them particularly suitable for processing time-resolved, sparse binary data. In this paper, we investigate the potential of leveraging SNNs for the detection of photon coincidences in positron emission tomography (PET) data. PET is a medical imaging technique based on injecting a patient with a radioactive tracer and detecting the emitted photons. One central post-processing task for inferring an image of the tracer distribution is the filtering of invalid hits occurring due to e.g. absorption or scattering processes. Our approach, coined PETNet, interprets the detector hits as a binary-valued spike train and learns to identify photon coincidence pairs in a supervised manner. We introduce a dedicated multi-objective loss function and demonstrate the effects of explicitly modeling the detector geometry on simulation data for two use-cases. Our results show that PETNet can outperform the state-of-the-art classical algorithm with a maximal coincidence detection F1 of 95.2%. At the same time, PETNet is able to predict photon coincidences up to 36 times faster than the classical approach, highlighting the great potential of SNNs in particle physics applications.

关键词： Spiking neural networks positron emission tomography particle coincidence detection supervised denoising

来源：评论

学校读者我要写书评

暂无评论

Convolution-Enhanced Evolving Attention networks

引用

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 2023年第7期45卷 8176-8192页

作者： Wang, Yujing Yang, Yaming Li, Zhuo Bai, Jiangang Zhang, Mingliang Li, Xiangtai Yu, Jing Zhang, Ce Huang, Gao Tong, Yunhai Peking Univ Key Lab Machine Percept MOE Sch Intelligence Sci & Technol Beijing 100871 Peoples R China Peking Univ Dept Comp Sci & Technol Engn Res Ctr Microprocessor & Syst Beijing 100871 Peoples R China Chinese Acad Sci Inst Informat Engn Beijing 100045 Peoples R China Swiss Fed Inst Technol CH-8092 Zurich Switzerland Tsinghua Univ Dept Automat Beijing 100190 Peoples R China

Attention-based neural networks, such as Transformers, have become ubiquitous in numerous applications, including computer vision, natural language processing, and time-series analysis. In all kinds of attention networks, the attention maps are crucial as they encode semantic dependencies between input tokens. However, most existing attention networks perform modeling or reasoning based on representations, wherein the attention maps of different layers are learned separately without explicit interactions. In this paper, we propose a novel and generic evolving attention mechanism, which directly models the evolution of inter-token relationships through a chain of residual convolutional modules. The major motivations are twofold. On the one hand, the attention maps in different layers share transferable knowledge, thus adding a residual connection can facilitate the information flow of inter-token relationships across layers. On the other hand, there is naturally an evolutionary trend among attention maps at different abstraction levels, so it is beneficial to exploit a dedicated convolution-based module to capture this process. Equipped with the proposed mechanism, the convolution-enhanced evolving attention networks achieve superior performance in various applications, including time-series representation, natural language understanding, machine translation, and image classification. Especially on time-series representation tasks, Evolving Attention-enhanced Dilated Convolutional (EA-DC-) Transformer outperforms state-of-the-art models significantly, achieving an average of 17% improvement compared to the best SOTA. To the best of our knowledge, this is the firstwork that explicitly models the layer-wise evolution of attention maps. Our implementation is available at https://***/pkuyym/EvolvingAttention.

关键词： Transformers Task analysis Convolution neural networks Machine translation Time series analysis Computational modeling Evolving attention image classification machine translation network architecture natural language understanding representation learning time series

来源：评论

学校读者我要写书评

暂无评论

DNCS: artificial Intelligence Enabled Food Recognition and Associated Calorie Level Estimation System Using Deep neural Classification Scheme

DNCS: Artificial Intelligence Enabled Food Recognition and A...

引用

2024 International Conference on Innovative Computing, Intelligent Communication and Smart Electrical Systems, ICSES 2024

作者： Ramesh, A.P. Prasad, Ovss Venkat, S. Jagadeeshwaran, A. Lakshmi, P. Adi Singaravelu, R. Sri Ramakrishna College of Arts and Science Department of Electronics Coimbatore India Qis College of Engineering and Technology Department of Ece AndhraPradesh Ongole India S.A. Engineering College Department of Ece Chennai India Sona College of Technology Department of Eee Salem India Qis College of Engineering and Technology Department of Cse AndhraPradesh Ongole India Vel Tech Multi Tech Dr. Rangarajan Dr Sakunthala Engineering College Department of Ece Chennai India

ISBN: (纸本)9798331543617

The DNCS (Cascaded CNN-ANN) enabled Food Recognition (FR) and Associated Calorie Level Estimation System presents a novel approach to food classification and calorie estimation using a cascaded neural network architecture. The system combines Convolutional neural networks (CNN) for food image recognition with artificial neural networks (ANN) for calorie estimation, leveraging fine-tuned pre-trained models such as InceptionV3 and EfficientNet. The system demonstrated high FR accuracy across various datasets, achieving a peak accuracy of 96.2% for common food items like burgers and pizzas, with an overall accuracy range between 89.2% and 96.2%. For calorie estimation, the system achieved a Mean Absolute Error (MAE) of less than 6.1%, demonstrating reliability in predicting calorie content based on food portion size. The use of image segmentation techniques like graph-cut and GrabCut further enhanced the accuracy of FR. The system was tested for real-time applications, achieving an average processing time of 350 to 450 milliseconds per image, making it suitable for integration into mobile applications. User feedback highlighted ease of use, speed, and accuracy as key strengths, with an overall satisfaction rating of 4.7 out of 5. These results suggest that the DNCS system offers an efficient solution for dietary management through real-time calorie tracking. © 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

Involution fused convolution for classifying eye-tracking patterns of children with Autism Spectrum Disorder

引用

ENGINEERING applications OF artificial INTELLIGENCE 2025年 139卷

作者： Islam, Md. Farhadul Manab, Meem Arafat Mondal, Joyanta Jyoti Zabeen, Sarah Bin Rahman, Fardin Hasan, Md. Zahidul Sadeque, Farig Noor, Jannatun BRAC Univ Sch Data & Sci Dept Comp Sci & Engn Comp Sustainabil & Social Good Res Grp C2SG Dhaka 1212 Bangladesh BRAC Univ Sch Data & Sci Dept Comp Sci & Engn Dhaka 1212 Bangladesh BRAC Univ Sch Data & Sci Dept Math & Nat Sci Dhaka 1212 Bangladesh Dublin City Univ Sch Law & Govt Dublin D09 V209 Ireland Univ Alabama Birmingham Dept Comp Sci Birmingham AL 35294 USA Concordia Univ Concordia Inst Informat Syst Engn CIISE Gina Cody Sch Engn & Comp Sci Montreal PQ H3G 1M8 Canada

Autism Spectrum Disorder (ASD) is a neurological condition that is challenging to diagnose. Numerous studies demonstrate that children diagnosed with autism struggle with maintaining attention spans and have less focused vision. The eye-tracking technology has drawn special attention in the context of ASD since anomalies in gaze have long been acknowledged as a defining feature of autism. Deep Learning (DL) approaches coupled with eye-tracking sensors are exploiting additional capabilities to advance the diagnostic and its applications. DL architectures like convolution have been dominating this domain. However, convolutions alone maybe insufficient to capture the important spatial information in eye-tracking patterns, as these patterns are more likely to be localized. The dynamic kernel-based process known as involutions can improve the classification efficiency. In this study, we utilize two image-processing operations to see how these processes learn eye-tracking patterns. Since these patterns are primarily based on spatial information, we employ a hybrid of involution and convolution. Our study shows that adding a few involution layers reduces size and computational cost while enhancing location-specific capabilities, maintaining performance comparable to pure convolutional models. However, excessive involution layers lead to weaker performance. For comparisons, we experiment with two datasets and a combined version of both in ablation studies. Our proposed model, featuring three involution layers and three convolution layers, achieved 99.43% accuracy on the first dataset and 96.78% on the second, with a size of only 1.36 megabytes. These results showcase the effectiveness of combining involution and convolution layers which outperforms previous literature.

关键词： Autism Spectrum Disorder Involutional neural networks Convolutional neural networks Eye-tracking Deep learning

来源：评论

学校读者我要写书评

暂无评论

DeepMesh: Differentiable Iso-Surface Extraction

引用

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 2024年第11期46卷 7072-7087页

作者： Guillard, Benoit Remelli, Edoardo Lukoianov, Artem Yvernay, Pierre Richter, Stephan R. Bagautdinov, Timur Baque, Pierre Fua, Pascal Ecole Polytech Fed Lausanne Comp Vis Lab CH-1015 Lausanne Switzerland MIT Grp Geometr Data Proc Cambridge MA 02139 USA Neural Concept CH-1015 Lausanne Switzerland Intel Labs Hillsboro OR 97124 USA Meta Real Labs Menlo Pk CA 94025 USA

Geometric Deep Learning has recently made striking progress with the advent of continuous deep implicit fields. They allow for detailed modeling of watertight surfaces of arbitrary topology while not relying on a 3D euclidean grid, resulting in a learnable parameterization that is unlimited in resolution. Unfortunately, these methods are often unsuitable for applications that require an explicit mesh-based surface representation because converting an implicit field to such a representation relies on the Marching Cubes algorithm, which cannot be differentiated with respect to the underlying implicit field. In this work, we remove this limitation and introduce a differentiable way to produce explicit surface mesh representations from Deep Implicit Fields. Our key insight is that by reasoning on how implicit field perturbations impact local surface geometry, one can ultimately differentiate the 3D location of surface samples with respect to the underlying deep implicit field. We exploit this to define DeepMesh - an end-to-end differentiable mesh representation that can vary its topology. We validate our theoretical insight through several applications: Single view 3D Reconstruction via Differentiable Rendering, Physically-Driven Shape Optimization, Full Scene 3D Reconstruction from Scans and End-to-End Training. In all cases our end-to-end differentiable parameterization gives us an edge over state-of-the-art algorithms.

关键词： Three-dimensional displays Topology Task analysis Surface reconstruction Shape Mesh generation Surface treatment Computational and artificial intelligence neural networks electronic design automation and methodology design methodology graphics shape computers and information processing image processing machine vision object recognition mathematics geometry computational geometry geometric modeling mathematics numerical analysis surface fitting systems engineering and theory systems, man, and cybernetics user interfaces data visualization isosurfaces

来源：评论

学校读者我要写书评

暂无评论

The local ternary pattern encoder-decoder neural network for dental image segmentation

引用

IET image processing 2022年第6期16卷 1520-1530页

作者： Salih, Omran Duffy, Kevin Jan Durban Univ Technol Inst Syst Sci ZA-4000 Durban South Africa

Recent advances in medical imaging analyses, especially the use of deep learning, are helping to identify, detect, classify, and quantify patterns in radiographs. At the centre of these advances is the ability to explore hierarchical feature representations learned from data. Deep learning is invaluably becoming the most sought out technique, leading to enhanced performances in the analysis of medical applications and systems. Deep learning techniques have achieved improved performance results in dental image segmentation. Segmentation of dental radiographs is a crucial step that helps dentists to diagnose dental caries. However, the performance of the deep networks used for these analyses are restrained by various challenging features found in dental carious lesions. Segmentation of dental images is often difficult due to the vast variety of types of topology, intricacies of medical structure and poor image quality caused by conditions such as low contrast, noise, irregular, and fuzzy border edges. These issues are exacerbated by low numbers of data images available for any particular analysis. A robust local ternary pattern encoder-decoder network (LTPEDN) is proposed to overcome dental image segmentation challenges and minimise the computational resources required. This new architecture is a modification of existing methods using an LTP. images are preprocessed via augmentation and normalisation techniques to increase and prepare the datasets. Thereafter, the dataset input is sent to the LTPEDN for training and testing the model. Segmentation is performed using the non-learnable layers (the LTP layers) and the learnable layers (standard convolution layers), to extract the ROI of the teeth. The method was evaluated on an augmented dataset of 11, 000 dental images. It was trained on 8, 800 training set images and tested on 2, 200 testing set images. The new method is shown to be 94.32% accurate.

关键词： dental image segmentation challenges deep networks convolutional neural nets deep learning techniques learning (artificial intelligence) hierarchical feature representations dentistry feature extraction medical structure local ternary pattern encoder-decoder neural network deep learning (artificial intelligence) image segmentation dental radiographs robust local ternary pattern encoder-decoder network data images dental caries medical imaging analyses enhanced performances diagnostic radiography medical image processing dental carious lesions improved performance results 000 dental images image classification medical applications

来源：评论

学校读者我要写书评

暂无评论

COMPREHENSIVE ANALYSIS OF CONVOLUTIONAL neural networks APPLIED TO CIFAR-10 DATASET 2

COMPREHENSIVE ANALYSIS OF CONVOLUTIONAL NEURAL NETWORKS APPL...

引用

2nd International Conference on Mechatronic Automation and Electrical Engineering, ICMAEE 2024

作者： Su, Biao College of Physics and Optoelectronic Engineering Shenzhen University Shenzhen518060 China

ISBN: (纸本)9781837242672

This article explores the advancements and applications of Convolutional neural networks (CNNs) in image classification, focusing on the CIFAR-10 dataset. Since their inception in 2006, CNNs have revolutionized computer vision by efficiently extracting high-level features from raw images. This study discusses the structure of CNNs, including convolutional layers, pooling, and fully connected layers, which enhance feature extraction and classification accuracy. Experimentally, CNNs have demonstrated superior performance on tasks requiring the identification of complex patterns such as in autonomous driving, security, and medical imaging. The research further investigates the optimization of CNN models through architectural enhancements, attention mechanisms, data augmentation, model compression, and cross-domain knowledge transfer. This study concludes that multicore configurations with batch processing significantly outperform single-core setups, achieving lower latency and higher throughput, thus underscoring the potent applications and continuous evolution of deep learning in modern artificial intelligence challenges. © The Institution of Engineering & Technology 2024.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Research on Crane Scene Recognition Technology Combining Digital Twins with neural networks 10

Research on Crane Scene Recognition Technology Combining Dig...

引用

10th International Conference on Fuzzy Systems and Data Mining, FSDM 2024

作者： Su, Wensheng Xu, Jia Wang, Xinren Special Equipment Safety Supervision Inspection Institute of Jiangsu Province Nanjing210036 China

ISBN: (纸本)9781643685694

In the fields of computer vision and intelligent recognition, with the increasing complexity and dynamism of crane scenes, how to use digital twins technology and artificial intelligence technology to improve the accuracy and efficiency of crane image recognition has become a research hotspot. Based on the combination of digital twin technology, convolutional neural networks, and long short-term memory networks (DTE-CNN-LSTM), an algorithm has been developed to achieve intelligent recognition and semantic understanding of complex crane operation scenes by introducing a virtual simulation environment for image feature optimization and processing. The research results indicate that the DTE-CNN-LSTM algorithm performs well in multiple crane scenarios. After optimization, the recognition accuracy of night cranes reached 100%, the recognition accuracy of rainy scenes increased to 95%, and the misclassification rate decreased to 5%. The recognition accuracy of obstacle has reached 98%, and the recognition accuracy of crane operation area has reached 90%, significantly reducing the misclassification rate. Overall, the DTE-CNN-LSTM algorithm achieved an accuracy rate of 96.70% and an F1 value of 96.64% in classification tasks for different crane scenarios. Its robustness and generalization ability in complex crane working environments have been verified, demonstrating high potential for application. © 2024 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

SVM-based malware classification through deep learning method 5

SVM-based malware classification through deep learning metho...

引用

5th International Conference on Signal processing and Computer Science, SPCS 2024

作者： Shi, Jiawei Wang, Wancheng Zhang, Meng Yao, Qinbo Cheng, Shiyuan Zhan, Zhifei College of Artificial Intelligence and Automation Hohai University Nanjing China

ISBN: (数字)9781510686731

ISBN: (纸本)9781510686724

Various daily applications, including image categorization, natural language comprehension, and voice identification, heavily rely on fully connected and convolutional neural networks. When tackling classification problems, conventional deep learning architectures typically employ softmax activation functions to normalize outputs and minimize the model's cross-entropy loss. In this study, we present a novel malware classification model that integrates hybrid support vector machines with neural networks. Notably, this model substitutes the conventional softmax layer in neural networks with a support vector machine, thereby shifting the learning process towards minimizing margin losses instead of cross-entropy losses. This modification enhances the precision of malware classification for both standard machine learning algorithms and several prevalent deep learning models. To validate our proposed model, we employed the Malimg dataset, comprising malware images derived from binary malware samples. Subsequently, we trained an DL-SVM model to assign scores to each distinct malware family. Experimental outcomes revealed that the malware classification model achieved an accuracy of 96.31%, a precision rate of 97.00%, a recall rate of 97.35%, and an F1 score of 97.16%, respectively. © 2025 SPIE.

关键词： Convolutional neural networks

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：