检索结果-内蒙古大学图书馆

11th Pacific-Rim Symposium on image and Video Technology, PSIVT 2023

ISBN: (纸本)9789819703753

The proceedings contain 34 papers. The special focus in this conference is on image and Video Technology. The topics include: Spatial Variation Sequences for Remote Sensing applications with Small Sample Sizes;exploring the Potential of High-Resolution Drone imagery for Improved 3D Human Avatar Reconstruction: A Comparative Study with Mobile images;point Cloud Novelty Detection Based on Latent Representations of a General Feature Extractor;Efficient 3Dconv Fusion of RGB and Optical Flow for Dynamic Hand Gesture Recognition and Localization;an Investigation of Video vision Transformers for Depression Severity Estimation from Facial Video Data;real-Time Automated Body Condition Scoring of Dairy Cows;Logo-SSL: Self-supervised Learning with Self-attention for Efficient Logo Detection;HAHANet: Towards Accurate image Classifiers with Less Parameters;evaluating Mammogram image Classification: Impact of Model Architectures, Pretraining, and Finetuning;melanoma Classification Using Deep Learning;3D Formation Control of Multiple Cooperating Autonomous Agents via Leader-Follower Strategy;LAPRNet: Lightweight Airborne Particle Removal Network for LiDAR Point Clouds;REAL-NET: A Monochromatic Depth Estimation Using REgional Attention and Local Feature Mapping;Spike-EFI: Spiking Neural Network for Event-Based Video Frame Interpolation;scrambleMix: A Privacy-Preserving image processing for Edge-Cloud machine Learning;Comparison of Simplified SE-ResNet and SE-DenseNet for Micro-Expression Classification;facial Deepfake Detection Using Gaussian Processes;A Novel Steganography Scheme Using Logistic Map, BRISK Descriptor, and K-Means Clustering;a Holistic Approach to Elderly Safety: Sensor Fusion, Fall Detection, and Privacy-Preserving Techniques;cluster-Based Video Summarization with Temporal Context Awareness;On Deploying Mobile Deep Learning to Segment COVID-19 PCR Test Tube images;enhancing Safety During Surgical Procedures with Computer vision, Artificial Intelligence, and Natural

关键词：

来源：评论

学校读者我要写书评

暂无评论

Deepfake attribution: On the source identification of artificially generated images

引用

WILEY INTERDISCIPLINARY REVIEWS-DATA MINING AND KNOWLEDGE DISCOVERY 2022年第3期12卷 e1438-e1438页

作者： Khoo, Brandon Phan, Raphael C. -W. Lim, Chern-Hong Monash Univ Sch IT Malaysia Campus Subang Jaya Malaysia Monash Univ Fac IT Dept Software Syst & Cybersecur Melbourne Vic Australia

Synthetic media or "deepfakes" are making great advances in visual quality, diversity, and verisimilitude, empowered by large-scale publicly accessible datasets and rapid technical progress in deep generative modeling. Heralding a paradigm shift in how online content is trusted, researchers in digital image forensics have responded with different proposals to reliably detect AI-generated images in the wild. However, binary classification of image authenticity is insufficient to regulate the ethical usage of deepfake technology as new applications are developed. This article provides an overview of the major innovations in synthetic forgery detection as of 2020, while highlighting the recent shift in research towards ways to attribute AI-generated images to their generative sources with evidence. We define the various categories of deepfakes in existence, the subtle processing traces and fingerprints that distinguish AI-generated images from reality and each other, and the different degrees of attribution possible with current understanding of generative algorithms. Additionally, we describe the limitations of synthetic image recognition methods in practice, the counter-forensic attacks devised to exploit these limitations, and directions for new research to assure the long-term relevance of deepfake forensics. Reliable, explainable, and generalizable attribution methods would hold malicious users accountable for AI-enabled disinformation, grant plausible deniability to appropriate users, and facilitate intellectual property protection of deepfake technology. This article is categorized under: Commercial, Legal, and Ethical Issues > Security and Privacy Algorithmic Development > Multimedia

关键词： computer vision machine learning deepfakes synthetic media source identification

来源：评论

学校读者我要写书评

暂无评论

Multi-modal and multi-dimensional biomedical image data analysis using deep learning

Multi-modal and multi-dimensional biomedical image data anal...

引用

作者： Wang, Yangyang University of Missouri

学位级别：博士

There is a growing need for the development of computational methods and tools for automated, objective, and quantitative analysis of biomedical signal and image data to facilitate disease and treatment monitoring, early diagnosis, and scientific discovery. Recent advances in artificial intelligence and machine learning, particularly in deep learning, have revolutionized computer vision and image analysis for many application areas. While processing of non-biomedical signal, image, and video data using deep learning methods has been very successful, high-stakes biomedical applications present unique challenges such as different image modalities, limited training data, need for explainability and interpretability etc. that need to be addressed. In this dissertation, we developed novel, explainable, and attention-based deep learning frameworks for objective, automated, and quantitative analysis of biomedical signal, image, and video data. The proposed solutions involve multi-scale signal analysis for oraldiadochokinesis studies; ensemble of deep learning cascades using global soft attention mechanisms for segmentation of meningeal vascular networks in confocal microscopy; spatial attention and spatio-temporal data fusion for detection of rare and short-term video events in laryngeal endoscopy videos; and a novel discrete Fourier transform driven class activation map for explainable-AI and weakly-supervised object localization and segmentation for detailed vocal fold motion analysis using laryngeal endoscopy videos. Experiments conducted on the proposed methods showed robust and promising results towards automated, objective, and quantitative analysis of biomedical data, that is of great value for potential early diagnosis and effective disease progress or treatment monitoring.

关键词：

来源：评论

学校读者我要写书评

暂无评论

FIRE-AD: frequency-dependent image reconstruction error for micro defect detection

引用

machine vision and applications 2025年第4期36卷 1-13页

作者： Nomura, Yuhei Hachiya, Hirotaka Department of Digital Manufacturing Industrial Technology Center of Wakayama Prefecture Wakayama-city Japan Graduate School of Systems Engineering Wakayama University Wakayama-city Japan

Micro defects, such as casting pores in industrial products, have been detected by human visual inspection using X-ray CT images and image processing tools. Although recent deep model-based methods achieve high anomaly detection performances, the detection of micro defects is challenging because metrics for anomaly detection are dominated by low-frequency information. To overcome the problem, we propose introducing frequency-dependent losses to capture reconstruction errors appearing around micro defects and frequency-dependent data augmentation to improve the sensitivity against the errors. We demonstrate the effectiveness of the proposed method through experiments with MVTec AD dataset especially on the detection of micro defects.

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

CNN-based On-board Intelligent processing for Remote Sensing images Interpretation

CNN-based On-board Intelligent Processing for Remote Sensing...

引用

image processing, Computer vision and machine Learning (ICICML), International Conference on

作者： Xiaojie Cui Yuanyuan Liu Xuehua Chen Gang Wang Beijing Institute of Remote Sensing Information Beijing China

Traditional remote sensing image processing is not able to provide timely information for near real-time applications due to the hysteresis of satellite-ground mutual communication and low processing efficiency. On-board intelligent processing is an important approach to improve the efficiency and intelligence of remote sensing satellites. This paper takes convolutional neural network (CNN) based on-board processing as the focus. Firstly, the basic workflow of CNN based on-board processing system is illustrated. Afterwards, the applications of lightweight CNN based on-board processing are thoroughly reviewed. The used CNN models are further analyzed to compare the advantages and disadvantages. Finally, current challenges are summarized and future works concerned with artificial intelligence are concluded.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SRUSTHI: super local bubble tracking inspired by machine vision

SRUSTHI: super local bubble tracking inspired by machine vis...

引用

IEEE International Ultrasonics Symposium (IUS)

作者： Sripada, Siva Saket Porter, Tyrone M. Univ Texas Austin Dept Biomed Engn Austin TX 78712 USA

ISBN: (数字)9781665466578

ISBN: (纸本)9781665466578

Super-resolution (SR) is a fascinating frontier in medical ultrasound (US) imaging offering the possibility of studying biological activity at spatiotemporal scales beyond the classical diffraction limit [1]. The key to SR is reliable detection and subsequent tracking of centroids of US contrast agents, over thousands of frames [1]. However, methods to overcome motion artefacts and background tissue speckle impose computational overhead [2];in addition to physical tradeoffs in data acquisition [1][3];thereby limiting biological applications to larger vessels with high blood flow rates [1]. The real-time or online nature of ultrasound imaging is sacrificed due to the offline nature of super-resolution processing methods [1]. In this work, we explore combinations of current machine vision algorithms, popular for similar object detection and tracking problems in optical imaging [4] - towards near real-time [5] super-resolution ultrasound imaging. We report encouraging results motivating further work towards improving state-of-the-art machine vision models designed for online, real-time, detection and tracking for ultrasound super-resolution.

关键词： biomedical ultrasound super-resolution computer vision machine learning image processing real-time

来源：评论

学校读者我要写书评

暂无评论

A novel opto-tactile sensing approach to enhance the handling of soft fruit

引用

COMPUTERS AND ELECTRONICS IN AGRICULTURE 2025年 235卷

作者： Ameur, Mohamed Adlan Ait El-Sayed, Amr M. Yan, Xiu T. Mehnen, Jorn Maier, Anja M. Univ Strathclyde Design Mfg & Engn Management Glasgow Scotland Assiut Univ Fac Engn Mechatron Engn Dept Assiut 71516 Egypt Glasgow Caledonian Univ Dept Mech Engn Cowcaddens Rd Glasgow G4 0BA Scotland

In agricultural settings, handling of soft fruit is critical to ensuring quality and safety. This study introduces a novel opto-tactile sensing approach designed to enhance the handling and assessment of soft fruit, with a case example of strawberries. Our approach utilises a Robotiq 2F-85 gripper equipped with the DIGIT vision-Based Tactile Sensor (VBTS) and attached to a Universal Robot UR10e. In contrast to force-based approaches, we introduce a novel purely image-based processing software pipeline for quantifying localised surface deformations in soft fruit. The system integrates fast and explainable image processing techniques applying image differencing, denoising, K-means clustering for unsupervised classification, morphological operations, and connected components analysis (CCA) to quantify surface deformations accurately. A calibration of the image processing pipeline using a rubber ball showed that the system effectively captured and analysed the rubber ball's surface deformations, benefiting from its uniform elasticity and predictable response to compression. As a soft fruit case example, the image processing pipeline was subsequently applied to strawberries, blueberries, and raspberries, demonstrating that the calibration parameters derived from the rubber ball could effectively assess surface deformations in soft fruits. Despite the complex, nonlinear deformation characteristics inherent to strawberries, blueberries, and raspberries, the pipeline exhibited robust performance, capturing and quantifying subtle surface changes. These findings underscore the system's capacity for precise deformation analysis in delicate materials, offering major potential for further refinement and adaptation. This novel approach of proposing and testing an image processing pipeline lays the groundwork for enhancing the handling and assessment of materials with intricate mechanical properties, paving the way for broader applications in sensitive agricultural and industrial

关键词： Robotic gripper vision-Based Tactile Sensor (VBTS) image processing machine Learning Soft Fruit Handling Robotic manipulation

来源：评论

学校读者我要写书评

暂无评论

An evaluation of platforms for processing camera-trap data using artificial intelligence

引用

METHODS IN ECOLOGY AND EVOLUTION 2023年第2期14卷 459-477页

作者： Velez, Juliana McShea, William Shamon, Hila Castiblanco-Camacho, Paula J. Tabak, Michael A. Chalmers, Carl Fergus, Paul Fieberg, John Univ Minnesota Dept Fisheries Wildlife & Conservat Biol St Paul MN 55108 USA Smithsonians Natl Zoo & Conservat Biol Inst Conservat Ecol Ctr Front Royal VA 22630 USA Univ Los Andes Dept Ciencias Biol Bogots Colombia Quantitat Sci Consulting LLC Toronto ON Canada Liverpool John Moores Univ Sch Comp Sci & Math Liverpool Merseyside England

Camera traps have quickly transformed the way in which many ecologists study the distribution of wildlife species, their activity patterns and interactions among members of the same ecological community. Although they provide a cost-effective method for monitoring multiple species over large spatial and temporal scales, the time required to process the data can limit the efficiency of camera-trap surveys. Thus, there has been considerable attention given to the use of artificial intelligence (AI), specifically deep learning, to help process camera-trap data. Using deep learning for these applications involves training algorithms, such as convolutional neural networks (CNNs), to use particular features in the camera-trap images to automatically detect objects (e.g. animals, humans, vehicles) and to classify species. To help overcome the technical challenges associated with training CNNs, several research communities have recently developed platforms that incorporate deep learning in easy-to-use interfaces. We review key characteristics of four AI platforms-Conservation AI, MegaDetector, MLWIC2: machine Learning for Wildlife image Classification and Wildlife Insights-and two auxiliary platforms-Camelot and Timelapse-that incorporate AI output for processing camera-trap data. We compare their software and programming requirements, AI features, data management tools and output format. We also provide R code and data from our own work to demonstrate how users can evaluate model performance. We found that species classifications from Conservation AI, MLWIC2 and Wildlife Insights generally had low to moderate recall. Yet, the precision for some species and higher taxonomic groups was high, and MegaDetector and MLWIC2 had high precision and recall when classifying images as either 'blank' or 'animal'. These results suggest that most users will need to review AI predictions, but that AI platforms can improve efficiency of camera-trap-data processing by allowing users to filt

关键词： artificial intelligence camera traps computer vision data processing deep learning image classification remote sensing review

来源：评论

学校读者我要写书评

暂无评论

iWBFCxr: improved weighted box fusion for analyzing chest x ray images for enhancing thoracic abnormalities

引用

International Journal of Information Technology (Singapore) 2025年第2期17卷 711-720页

作者： Verma, Sonia Devarajan, Ganesh Gopal Sharma, Pankaj Kumar Department of Computer Science and Engineering Faculty of Engineering and Technology SRM Institute of Science and Technology Delhi - NCR Campus Delhi - Meerut Road Modinagar Uttar Pradesh Ghaziabad 201204 India Department of Computer Science ABES Engineering College UP Ghaziabad 201009 India

Chest x-ray studies can be automatically detected and their locations located using artificial intelligence (AI) in healthcare. To detect the location of findings, additional annotation in the form of bounding boxes is required, rather than image-level labeling. Accurate interpretation of chest radiographs is essential for diagnosing thoracic abnormalities. This paper introduces a novel method for abnormality detection in chest radiographs using the VinDr-CXR dataset. Labeled training sets are improved by incorporating data augmentation and advanced pre-processing techniques. To improve the accuracy of anomaly identification, the study makes use of ensemble object detection models, such as the modified weighted boxes fusion (iWBFCxr) technique. With an average precision (AP) of 0.2434 for class 14, experimental results on the VinDr-CXR dataset show an overall mean average precision (mAP) of 0.2791. The iWBFCxr approach, when combined with ensemble predictions, has the potential to improve abnormality localization. By integrating enhanced label merging with ensemble object detection models, our method improves abnormality localization in chest radiographs significantly, improving thoracic abnormality diagnosis accuracy and possibly leading to new medical interventions. © Bharati Vidyapeeth's Institute of Computer applications and Management 2024.

关键词： Automated classification Automated localization Chest radiographs Computer vision Deep learning machine learning Thoracic anomalies

来源：评论

学校读者我要写书评

暂无评论

Advancements and Challenges in Generative AI: Architectures, applications, and Ethical Implications

Advancements and Challenges in Generative AI: Architectures,...

引用

2024 Ital-IA Intelligenza Artificiale - Thematic Workshops, Ital-IA 2024

作者： Amato, Flora Benfenati, Domenico Cirillo, Egidia De Filippis, Giovanni Maria Fonisto, Mattia Galli, Antonio Marrone, Stefano Marassi, Lidia Moscato, Vincenzo Patwardhan, Narendra Moccardi, Alberto Pascarella, Antonio Elia Rinaldi, Antonio M. Russo, Cristiano Sansone, Carlo Tommasino, Cristian University of Naples Federico II Via Claudio 21 Naples80125 Italy University of Naples Federico II Naples Italy

Architecture, classification, and major applications of Generative AI interfaces, specifically chatbots, are presented in this paper. Research paper details how the Generative AI interfaces work with various Generative AI approaches and show the architecture and their working. On the other hand, the generative model is built using advanced machine learning techniques to build dynamic, contextually relevant responses automatically. On the other hand, the retrieval-based model builds up with dependency on a predefined response library. The paper also discusses the use of Generative AI to populate Multimedia Knowledge Graphs (KGs), presenting technologies based on the semantic analysis of deep learning and NoSQL to more effectively integrate and retrieve data. The social and ethical challenges that come with the deployment of generative models are critically reviewed. These dialogues bring forward the balance that has to be maintained between progress and necessity in technological advancements, for which the call for ethical responsibility in developing AI is made. The paper presents a comprehensive review of state-of-the-art Generative AI with special focus on the promises and pitfalls in Generative AI research related to both natural language processing and knowledge management. © 2024 Copyright for this paper by its authors.

关键词： Generative adversarial 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：