检索结果-内蒙古大学图书馆

2nd IEEE World Conference on Communication and Computing (WCONF)

作者： Chinni, Naga Praneeth Kumar Kaamaala, Sai Pranav Reddy Vardhan, Bhasham Vishva Kishan, Adari Uday Richards, Vasimalla Sunny Vardhan, Ramadasu Nooka Harsh Puneet Lovely Profess Univ Comp Sci & Engn Jalandhar Punjab India

ISBN: (纸本)9798350395334;9798350395327

This project introduces a transformative object detection system designed to enhance the navigational capabilities of visually impaired individuals through the application of advanced computer vision technologies. Utilizing the You Only Look Once (YOLO) model, paired with the Comprehensive Object Collection (COCO) dataset, this system provides real-time, accurate object detection and classification. The core functionality of the application allows for the processing of both static images and live video feeds, enabling blind users to receive auditory announcements of nearby objects, thereby assisting with spatial awareness and environmental interaction. The system leverages a pre-trained YOLO model to ensure robust detection performance, achieving a peak detection accuracy of 99%. By delivering object labels and bounding box coordinates audibly, the application serves as a critical tool in improving the daily independence and quality of life for people with visual impairments. This project not only highlights the potential of deep learning in assistive technologies but also underscores the importance of adaptive solutions in inclusive technology development.

关键词： YOLO model COCO dataset object detection visually impaired assistance real-time processing audio feedback computer vision assistive technology

来源：评论

学校读者我要写书评

暂无评论

A novel deep learning-based technique for efficient characterization of engineered cementitious composites cracks for durability assessment

引用

STRUCTURAL CONCRETE 2025年第2期26卷 2107-2123页

作者： Das, Avik Kumar Leung, Christopher K. Y. Tsinghua Univ Inst Ocean Engn SIGS Shenzhen Peoples R China HKUST Dept Civil & Environm Engn Hong Kong Peoples R China

Engineered Cementitious Composites also known as Strain-hardening cementitious composites (SHCCs) has unique cracking patterns like cracks that have tiny widths and showcase high density. All of this makes it difficult and laborious to compute crack parameters from crack patterns. Unfortunately, this is an essential part of assessing durability and micromechanical modeling. SHSnet is developed to perform end-to-end semantic segmentation of SHCC cracks. SHSnet is efficient, attention based deep encoder-decoder network with large receptive field. Loss function based on Tversky function were used for training the model. SHSnet with loss function shows promising result with mPrecision, mF1Score and mIoU of 0.87, 0.84 and 0.83 respectively for complex SHCC cracks while requiring at least an order of fewer computational parameters than those in the literature. An image processing unit is then used to estimate the width, number, and length of the cracks from the segmentation mask. Test results show that the computed crack parameters with SHSnet are exactly the same as that computed with an optical microscope but require similar to 100x less time. Results demonstrate that SHSnet works equally well in SHCCs with different surface textures, crack density, and widths;the final result was far superior to a conventional technique. This technique also shows promising results in an automatic evaluation of crack parameters relevant to durability and visualizing crack patterns even in the presence of artifacts during progressive testing. The results also demonstrate the necessity to accurately and densely calculate crack length and maximum crack width;else the durability results are expected to be significantly more conservative than the actual value.

关键词： deep learning engineered cementitious composites multiple-thin- tortuous crack characterization strain hardening cementitious composites

来源：评论

学校读者我要写书评

暂无评论

Generating high-resolution hyperspectral time series datasets based on unsupervised spatial-temporal-spectral fusion network incorporating a deep prior

引用

INFORMATION FUSION 2024年 111卷

作者： Sun, Weiwei Ren, Kai Meng, Xiangchao Yang, Gang Liu, Qiang Zhu, Lin Peng, Jiangtao Li, Jiancheng Ningbo Univ Dept Geog & Spatial Informat Tech Ningbo 315211 Peoples R China Ningbo Univ Fac Elect Engn & Comp Sci Ningbo 315211 Peoples R China Hubei Univ Fac Math & Stat Hubei Key Lab Appl Math Wuhan 430062 Peoples R China Wuhan Univ Sch Geodesy & Geomat Wuhan 430079 Peoples R China

Over the past decade, image fusion has emerged as an indispensable tool for surface monitoring due to its capability to reconstruct high-quality surface reflectance. While satisfactory progress has been made in the spatial-spectral and spatial-temporal fusion of remote sensing images, the fusion of spatial-temporal-spectral data remains a challenging task. This challenge arises from the presence of multiple nonlinear relationships caused by inconsistent temporal, spatial, and spectral resolutions. In this paper, we propose a novel approach to address the spatial-temporal-spectral fusion of remote sensing images using a triple discriminator Generative Adversarial Network called TDGAN. Our approach represents the first attempt at spatial-temporal-spectral fusion based on deep learning techniques. Specifically, TDGAN achieves unsupervised learning by incorporating a deep spectral transformation prior, explores the application performance of the Swin Transformer in spectral super-resolution for the first time. Additionally, three discriminators are designed based on the prior consistencies: phenological consistency for data acquired at different times and spatial-spectral consistency for data acquired at the same time. These discriminators effectively enhance the prediction accuracy of missing hyperspectral images. Extensive simulations and real experiments were conducted to evaluate the effectiveness of our proposed TDGAN. The results demonstrate its robustness and comparable performance to other state-of-the-art methods. The proposed approach provides high-quality hyperspectral datasets with high spatial and temporal resolutions, facilitating continuous land cover observations.

关键词： Spatial -temporal -spectral fusion Hyperspectral Multispectral TDGAN

来源：评论

学校读者我要写书评

暂无评论

FireLens: A deep learning Solution for Wildfire Identification 6

FireLens: A Deep Learning Solution for Wildfire Identificati...

引用

6th IEEE International Conference on Emerging Smart Computing and Informatics, ESCI 2024

作者： Pawar, Shital Salunkhe, Parag Jaybhay, Sarthak Vishwakarma Institute of Technology Dept. of Electronics and tele Pune India

ISBN: (纸本)9798350306613

This paper presents an optimized framework for video smoke detection tailored specifically for deployment on local cameras in embedded systems. The proposed methodology exhibits its effectiveness in real-time smoke image identification with minimal false positive occurrences, as substantiated through rigorous experimentation. By harnessing the capabilities of CNN technology, system showcases promising capabilities for swiftly detecting smoke associated with wildfires. The emphasis is placed on the importance of forest fire early warning systems prevention, and the design of this framework is geared towards promptly and accurately recognizing potential threats. The incorporation of deep learning in wildfire smoke detection offers a robust solution for proactive incident management, enabling swift responses to potential wildfire incidents. © 2024 IEEE.

关键词： Convolutional neural network smoke detection Wildfires

来源：评论

学校读者我要写书评

暂无评论

EmotionCast: An Emotion-Driven Intelligent Broadcasting System for Dynamic Camera Switching

引用

SENSORS 2024年第16期24卷 5401页

作者： Zhang, Xinyi Ba, Xinran Hu, Feng Yuan, Jin Commun Univ China Sch Informat & Commun Engn Beijing 100024 Peoples R China

Traditional broadcasting methods often result in fatigue and decision-making errors when dealing with complex and diverse live content. Current research on intelligent broadcasting primarily relies on preset rules and model-based decisions, which have limited capabilities for understanding emotional dynamics. To address these issues, this study proposed and developed an emotion-driven intelligent broadcasting system, EmotionCast, to enhance the efficiency of camera switching during live broadcasts through decisions based on multimodal emotion recognition technology. Initially, the system employs sensing technologies to collect real-time video and audio data from multiple cameras, utilizing deep learning algorithms to analyze facial expressions and vocal tone cues for emotion detection. Subsequently, the visual, audio, and textual analyses were integrated to generate an emotional score for each camera. Finally, the score for each camera shot at the current time point was calculated by combining the current emotion score with the optimal scores from the preceding time window. This approach ensured optimal camera switching, thereby enabling swift responses to emotional changes. EmotionCast can be applied in various sensing environments such as sports events, concerts, and large-scale performances. The experimental results demonstrate that EmotionCast excels in switching accuracy, emotional resonance, and audience satisfaction, significantly enhancing emotional engagement compared to traditional broadcasting methods.

关键词： emotion recognition intelligent broadcasting multimodal data processing

来源：评论

学校读者我要写书评

暂无评论

deep learning based fine-grained recognition technology for basketball movements

引用

SYSTEMS AND SOFT COMPUTING 2024年 6卷

作者： Zhang, Lin Jilin Prov Econ Management Cadre Coll Changchun 130012 Peoples R China

The technical essentials of basketball are very complex, in order to accomplish the identification of fine-grained basketball movements and improve the quality of basketball training, the study innovatively proposes a region screening strategy to obtain the local region information of image frames, and draws on the SlowFast network structure to design the static and dynamic detail models, and finally utilizes the 3D attention feature fusion module to complete the multi-feature relationship processing and recognition. Performance testing experiments have confirmed that the dual stream model based on region filtering has achieved a maximum improvement in average accuracy of 16.35 % and 60.36 % in the recognition of RGB images, preserving relatively complete local effective region information of RGB images. The model based on channel domain attention mechanism improves its average accuracy to over 60 %. The model recognition accuracy calculation time after dual stream fusion is reduced by about 50 %. The model based on 3D attention feature fusion module exhibits stronger robustness when the feature dimension changes. And the loss value of this model rapidly decreases and converges to the minimum loss value, with its accuracy and recall reaching 90.03 % and 88.47 % respectively, and its AUC value reaching a maximum of 0.761. The design of the study contributes to the digital development of teaching and training in basketball and enriches the theoretical body of knowledge on deep learning and action recognition.

关键词： Fine-grained Basketball Attention mechanism Multi scale features Three-dimensional convolution

来源：评论

学校读者我要写书评

暂无评论

Frame-Adaptive Multi-Object Tracking for Discontinuous image Sequences 2

Frame-Adaptive Multi-Object Tracking for Discontinuous Image...

引用

2nd International Conference on Algorithm, image processing and Machine Vision, AIPMV 2024

作者： Li, Xiaolei Xiao, Xingjie Yang, Siyuan Sha, Moquan Sha, Zongyao Tu, Jianguang School of Remote Sensing and Information Engineering Wuhan University Wuhan China Wuhan Luojia Zhongheng Remote Sensing Data Technology Co Wuhan China China Unicom Smart City Research Institute Beijing China

ISBN: (纸本)9798350390254

In the field of multi-object tracking, this study introduces an innovative framework designed to address the challenges posed by frame loss in image sequences, particularly within the contexts of video surveillance and traffic management. Despite the significant achievements of deep learning-based tracking methods in processing high frame rate videos, frame loss remains a prevalent issue in real-world deployments due to network and hardware constraints, which can result in erroneous target associations and interrupted trajectory tracking. Our research begins by detecting frame loss through a timestamp- based analysis, this is followed by the deployment of an adaptive algorithm that fuses spatial and appearance metrics to compensate for the positional discrepancies and visual inconsistencies caused by missing frames, thus improving the precision of data association. We have curated a dataset that accurately reflects real-world environments, and conducted validation on subsampled versions of the MOT17 and MOT20 datasets, in addition to our original dataset. Empirical results indicate that our multi-object tracking approach demonstrates significant adaptability and robustness in handling discontinuous image sequences, reducing the number of ID switches by 54.1 % on our original dataset compared to state-of-the-art methods. The proposed framework has the potential to significantly enhance the performance of multi-object tracking systems in real-world scenarios where frame loss is a common occurrence, thereby improving the reliability and accuracy of video surveillance and traffic management applications. © 2024 IEEE.

关键词： Adaptive algorithms

来源：评论

学校读者我要写书评

暂无评论

ULM-MbCNRT: In Vivo Ultrafast Ultrasound Localization Microscopy by Combining Multibranch CNN and Recursive Transformer

引用

IEEE TRANSACTIONS ON ULTRASONICS FERROELECTRICS AND FREQUENCY CONTROL 2024年第12期71卷 1735-1751页

作者： Zhang, Gaobo Gu, Wenting Yue, Yaoting Tang, Meng-Xing Luo, Jianwen Liu, Xin Ta, Dean Fudan Univ Dept Biomed Engn Shanghai Peoples R China Fudan Univ Acad Engn & Technol Shanghai Peoples R China Imperial Coll London Dept Bioengn London England Biomed Engn Tsinghua Univ Beijing Peoples R China Fudan Univ Dept Elect Engn Shanghai Peoples R China

Ultrasound localization microscopy (ULM) overcomes the acoustic diffraction limit by localizing tiny microbubbles (MBs), thus enabling the microvascular to be rendered at subwavelength resolution. Nevertheless, to obtain such superior spatial resolution, it is necessary to spend tens of seconds gathering numerous ultrasound (US) frames to accumulate the MB events required, resulting in ULM imaging still suffering from tradeoffs between imaging quality, data acquisition time, and data processing speed. In this article, we present a new deep learning (DL) framework combining multibranch convolutional neural network (CNN) and recursive transformer (RT), termed ULM-MbCNRT, that is capable of reconstructing a super-resolution (SR) image directly from a temporal mean low-resolution image generated by averaging much fewer raw US frames, i.e., implement an ultrafast ULM imaging. To evaluate the performance of ULM-MbCNRT, a series of numerical simulations and in vivo experiments are carried out. Numerical simulation results indicate that ULM-MbCNRT achieves high-quality ULM imaging with similar to 10-fold reduction in data acquisition time and similar to 130-fold reduction in computation time compared to the previous DL method (e.g., the modified subpixel CNN, ULM-mSPCN). For the in vivo experiments, when comparing to the ULM-mSPCN, ULM-MbCNRT allows similar to 37-fold reduction in data acquisition time (similar to 0.8 s) and similar to 2134-fold reduction in computation time (similar to 0.87 s) without sacrificing spatial resolution. It implies that ultrafast ULM imaging holds promise for observing rapid biological activity in vivo, potentially improving the diagnosis and monitoring of clinical conditions.

关键词： Imaging Transformers Location awareness Convolutional neural networks Spatial resolution Feature extraction Microscopy Convolutional neural network (CNN) deep learning (DL) super-resolution ultrasound (US) imaging transformer ultrasound localization microscopy (ULM)

来源：评论

学校读者我要写书评

暂无评论

Research on Computer deep learning Algorithm to Optimize the Fault Prediction Model of Nuclear Power Plant Emergency Feed Pump 2

Research on Computer Deep Learning Algorithm to Optimize the...

引用

2nd IEEE International Conference on image processing and Computer Applications, ICIPCA 2024

作者： Gao, Chao Song, Xianjun Zhang, Zhiqiang Li, Chunguang Chang, Xincai Cgn Digital Technology Co. Ltd Cgnpc Beijing China

ISBN: (纸本)9798350360240

This project takes the typical failure forms of pumps in nuclear power plants such as abnormal vibration, friction and wear as the research object. The most readily available pump housing acceleration signal frequency domain information is used as input. The system is mainly composed of four modules: input module, convolution module, multi-head self-attention module and output module. The attention mechanism of characteristic frequency domain data based on the fusion of deep neural network and attention network is studied. At the same time, the pattern recognition model of pump faults in nuclear power plant is constructed. Compared with the existing research results, using frequency-domain data as input, using frequency-domain data attention network and other technologies can shorten the amount of input data, and make the recognition accuracy of the model up to 100% in the test set, which has obvious advantages over other fault recognition models based on deep neural networks. © 2024 IEEE.

关键词： Nuclear energy

来源：评论

学校读者我要写书评

暂无评论

Molten Pool image Segmentation Based on Adaptive Multi-Scale Attention Mechanism

引用

IEEE ACCESS 2025年 13卷 73136-73145页

作者： Chen, Yuefeng Liu, Weihua Yang, Qing Chen, Huabin Jiang, Qi Shanghai Jiao Tong Univ Sch Mat Sci & Engn Shanghai 201500 Peoples R China China Nucl Ind Fifth Construct Co Ltd Shanghai 201500 Peoples R China Nanjing Univ Posts & Telecommun Coll Automat Nanjing 210000 Peoples R China Nanjing Univ Posts & Telecommun Coll Artificail Intelligence Nanjing 210000 Peoples R China

This study presents an innovative segmentation algorithm that seamlessly integrates an adaptive multi-scale attention mechanism with the Lossab loss function to tackle challenges like overlapping molten pool contours and arc interference during image segmentation. Initially, an adaptive multi-scale channel attention mechanism is introduced and fused with a feature pyramid, forming a multi-scale attention pyramid structure that significantly enhances the network's capability to extract intricate molten pool features. Subsequently, a tailored Lossab loss function is devised to boost the network's precision in detecting the edges of the molten pool. Experimental outcomes reveal that the proposed method attains a remarkable segmentation accuracy of 96.77% within the molten pool region on the TIG molten pool image dataset. The AMSNet algorithm outperforms PIDNet by 2.51% and SegNet by 4.66% in terms of molten pool segmentation accuracy, providing solid evidence of its effectiveness in tackling the task of molten pool image segmentation. The model performs well in computational efficiency, data processing and complexity balance, making it suitable for industrial field deployment.

关键词： Welding Feature extraction image segmentation Production Attention mechanisms Adaptation models Accuracy Cameras real-time systems Charge coupled devices Semantic segmentation molten pool images attention mechanism deep learning industrial vision

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：