检索结果-内蒙古大学图书馆

Fusion of Ship Perceptual Information for Electronic Navigational Chart and Radar images based on deep learning

JOURNAL OF NAVIGATION 2020年第1期73卷 192-211页

作者： Guo, Muzhuang Guo, Chen Zhang, Chuang Zhang, Daheng Gao, Zongjiang Dalian Maritime Univ Coll Marine Elect Engn Dalian 116026 Liaoning Peoples R China Dalian Maritime Univ Nav Coll Dalian 116026 Peoples R China

Superimposing Electronic Navigational Chart (ENC) data on marine radar images can enrich information for navigation. However, direct image superposition is affected by the performance of various instruments such as Global Navigation Satellite Systems (GNSS) and compasses and may undermine the effectiveness of the resulting information. We propose a data fusion algorithm based on deep learning to extract robust features from radar images. By deep learning in this context we mean employing a class of machine learning algorithms, including artificial neural networks, that use multiple layers to progressively extract higher level features from raw input. We first exploit the ability of deep learning to perform target detection for the identification of marine radar targets. Then, image processing is performed on the identified targets to determine reference points for consistent data fusion of ENC and marine radar information. Finally, a more intelligent fusion algorithm is built to merge the marine radar and electronic chart data according to the determined reference points. The proposed fusion is verified through simulations using ENC data and marine radar images from real ships in narrow waters over a continuous period. The results suggest a suitable performance for edge matching of the shoreline and real-time applicability. The fused image can provide comprehensive information to support navigation, thus enhancing important aspects such as safety.

关键词： deep learning Marine radar Electronic Navigational Chart (ENC) Data fusion

来源：评论

学校读者我要写书评

暂无评论

PROBABLY PLEASANT? A NEURAL-PROBABILISTIC APPROACH TO AUTOMATIC MASKER SELECTION FOR URBAN SOUNDSCAPE AUGMENTATION 47

PROBABLY PLEASANT? A NEURAL-PROBABILISTIC APPROACH TO AUTOMA...

引用

47th IEEE International Conference on Acoustics, Speech and Signal processing (ICASSP)

作者： Ooi, Kenneth Watcharasupat, Karn N. Lam, Bhan Ong, Zhen-Ting Gan, Woon-Seng Nanyang Technol Univ NTU Sch Elect & Elect Engn Singapore Singapore

ISBN: (纸本)9781665405409

Soundscape augmentation, which involves the addition of sounds known as "maskers" to a given soundscape, is a human-centric urban noise mitigation measure aimed at improving the overall soundscape quality. However, the choice of maskers is often predicated on laborious processes and is inflexible to the time-varying nature of real-world soundscapes. Owing to the perceptual uniqueness of each soundscape and the inherent subjectiveness of human perception, we propose a probabilistic perceptual attribute predictor (PPAP) that predicts parameters of random distributions as outputs instead of a single deterministic value. Using the PPAP, we developed a novel automatic masker selection system (AMSS), which selects optimal masker candidates based on the predicted distribution of the ISO 12913-3 Pleasantness score for a given soundscape. Via a largescale listening test with 300 participants, we collected 12600 subjective responses, each to a unique augmented soundscape, to train the PPAP models in a 5-fold cross-validation scheme. Using a convolutional recurrent neural network backbone and experimenting with several variants of the attention mechanism for the PPAP, we evaluated the proposed system on a blind test set with 48 unseen augmented soundscapes to assess the effectiveness of the probabilistic output scheme over traditional deterministic systems.

关键词： Soundscape neural attention soundscape augmentation deep learning probabilistic model

来源：评论

学校读者我要写书评

暂无评论

Single image Deraining Using time-Lapse Data

引用

IEEE TRANSACTIONS ON image processing 2020年 29卷 7274-7289页

作者： Cho, Jaehoon Kim, Seungryong Min, Dongbo Sohn, Kwanghoon Yonsei Univ Sch Elect & Elect Engn Seoul 120749 South Korea Ecole Polytech Fed Lausanne EPFL CH-1015 Lausanne Switzerland Korea Univ Dept Comp Sci & Engn Seoul 02841 South Korea Ewha Womans Univ Dept Comp Sci & Engn Seoul 03760 South Korea

Leveraging on recent advances in deep convolutional neural networks (CNNs), single image deraining has been studied as a learning task, achieving an outstanding performance over traditional hand-designed approaches. Current CNNs based deraining approaches adopt the supervised learning framework that uses a massive training data generated with synthetic rain streaks, having a limited generalization ability on real rainy images. To address this problem, we propose a novel learning framework for single image deraining that leverages time-lapse sequences instead of the synthetic image pairs. The deraining networks are trained using the time-lapse sequences in which both camera and scenes are static except for time-varying rain streaks. Specifically, we formulate a background consistency loss such that the deraining networks consistently generate the same derained images from the time-lapse sequences. We additionally introduce two loss functions, the structure similarity loss that encourages the derained image to be similar with an input rainy image and the directional gradient loss using the assumption that the estimated rain streaks are likely to be sparse and have dominant directions. To consider various rain conditions, we leverage a dynamic fusion module that effectively fuses multi-scale features. We also build a novel large-scale time-lapse dataset providing real world rainy images containing various rain conditions. Experiments demonstrate that the proposed method outperforms state-of-the-art techniques on synthetic and real rainy images both qualitatively and quantitatively. On the high-level vision tasks under severe rainy conditions, it has been shown that the proposed method can be utilized as a pre-preprocessing step for subsequent tasks.

关键词： Rain Training data Task analysis Convolutional neural networks Rendering (computer graphics) Training Feature extraction Single image deraining convolutional neural networks (CNNs) time-lapse dataset dynamic fusion module

来源：评论

学校读者我要写书评

暂无评论

RAINING DETECTION WITH deep learning METHOD FOR VEHICLE SYSTEM

RAINING DETECTION WITH DEEP LEARNING METHOD FOR VEHICLE SYST...

引用

IEEE International Conference on Consumer Electronics (ICCE)

作者： Zheng, Meng-Jie Hsia, Shih-Chang Wang, Szu-Hong Natl Yunlin Univ Sci & Technol Dept Elect Engn Touliu Taiwan

ISBN: (纸本)9781728197661

This research develops the raining detection using deep learning method. A training model is established through a neural network-like architecture to classify the images recorded by the vehicle camera into three classes: no rain, heavy rain, and light rain. The classification results are used to control the speed of the motor to implement an automatic wiper control system. The training model with ResNet is used to classify the image. For real-time application, the camera module on the Google Coral Dev board embedded development platform is used to test the video stream and to estimate the performance of this system. Results show that the recognition accuracy reaches 95%, and the processing speed can achieve 20 frames per second (fps).

关键词： Rain Detection Neural Network ADAS Embedded System

来源：评论

学校读者我要写书评

暂无评论

Automatic Classification System for Diagnosis of Cognitive Impairment Based on the Clock-Drawing Test 9th

Automatic Classification System for Diagnosis of Cognitive I...

引用

9th International Work-Conference on the Interplay Between Natural and Artificial Computation (IWINAC)

作者： Jimenez-Mesa, C. Arco, Juan E. Valenti-Soler, M. Frades-Payo, B. Zea-Sevilla, M. A. Ortiz, A. Avila-Villanueva, M. Castillo-Barnes, Diego Ramirez, J. Del Ser-Quijano, T. Carnero-Pardo, C. Gorriz, J. M. Data Sci & Computat Intelligence DASCI Inst Granada Spain Univ Granada Signal Theory Telemat & Commun Dept Granada Spain Univ Malaga Commun Engn Dept Malaga Spain Fdn CIEN Madrid Spain FIDYAN Neuroctr Granada Spain

ISBN: (纸本)9783031062421;9783031062414

The prevalence of dementia is currently increasing worldwide. This syndrome produces a deterioration in cognitive function that can not be reverted. However, an early diagnosis can be crucial for slowing its progress. The Clock Drawing Test (CDT) is a widely used paperand-pencil test for cognitive assessment in which an individual has to manually draw a clock on a paper during a certain time. Nevertheless, there are a lot of scoring systems for this test and most of them depend on the subjective assessment of the expert. This study proposes a computeraided diagnosis (CAD) system based on deep learning in order to automate the diagnosis of cognitive impairment (CI) from the result of the CDT. This is addressed by employing a preprocessing pipeline in which the clock is detected and centered, as well as binarized for decreasing the computational burden. Then, the resulting image is fed into a Convolutional Neural Network (CNN), which is used to identify the informative patterns within the CDT drawings that are relevant for the assessment of the patient's cognitive status. Performance is evaluated in a real context where differentiating between CI patients and controls. The proposed method provides an accuracy of 68.62% in this classification task, with an AUC of 74.53%. A validation method using resubstitution with upper bound correction is also discussed.

关键词： Alzheimer's disease Clock Drawing Test Cognitive impairment deep learning image processing Machine learning

来源：评论

学校读者我要写书评

暂无评论

deep learning-Based 3D Multi-Object Tracking Using Multimodal Fusion in Smart Cities

引用

HUMAN-CENTRIC COMPUTING AND INFORMATION SCIENCES 2024年 14卷

作者： Li, Hui Liu, Xiang Jia, Hong Ahanger, Tariq Ahamed Xu, Lingwei Alzamil, Zamil Li, Xingwang Qingdao Univ Sci & Technol Sch Informat Sci & Technol Qingdao Peoples R China Minist Educ Engn Res Ctr Integrat & Applicat Digital Learning Beijing Peoples R China Xiamen Univ Sch Informat Fujian Key Lab Sensing & Comp Smart Cities Xiamen Peoples R China Prince Sattam bin Abdulaziz Univ Coll Comp Engn & Sci Alkharj Saudi Arabia Majmaah Univ Coll Comp & Informat Sci Dept Comp Sci Al Majmaah Saudi Arabia Henan Polytech Univ Sch Phys & Elect Informat Engn Jiaozuo Peoples R China

The intelligent processing of visual perception information is one of the core technologies of smart cities. deep learning-based 3D multi-object tracking is important in improving the intelligence and safety of robots in smart cities. However, 3D multi-object tracking still faces many challenges due to the complexity of the environment and uncertainty of the object. In this paper, we make the most of the multimodal information of image and point cloud and propose a multimodal adaptive feature gating fusion module to improve the feature fusion effect. In the object association stage, we designed an orientation-position-aware affinity matrix (EO-IoU) by using Euclidean distance, orientation similarity, and intersection over union, which is more suitable for the association to solve the problem of association failure when there is little or no overlap between the detection box and the prediction box. At the same time, we adopt a more robust two-stage data association method to solve the trajectory fragmentation and identity switching caused by discarding low-scoring detection boxes. The results of extensive experiments on the KITTI and NuScenes benchmark datasets demonstrate that our method outperforms existing state-of-the-art methods with better robustness and accuracy.

关键词： Smart Cities Visual Perception 3D Multi-Object Tracking Multimodal Feature Fusion Position Affinity Matrix Data Association

来源：评论

学校读者我要写书评

暂无评论

deep learning-based masonry crack segmentation and real-life crack length measurement

引用

CONSTRUCTION AND BUILDING MATERIALS 2022年 359卷

作者： Dang, L. Minh Wang, Hanxiang Li, Yanfen Nguyen, Le Quan Nguyen, Tan N. Song, Hyoung-Kyu Moon, Hyeonjoon Sejong Univ Dept Comp Sci & Engn 209 Neungdong ro Seoul 05006 South Korea Sejong Univ Dept Informat & Commun Engn Seoul South Korea Sejong Univ Convergence Engn Intelligent Drone Seoul South Korea FPT Univ Dept Informat Technol Ho Chi Minh City 70000 Vietnam Sejong Univ Dept Architectural Engn 209 Neungdong ro Seoul 05006 South Korea

While there have been a considerable number of studies on computer vision (CV)-based crack detection on concrete/asphalt public facilities, such as sewers and tunnels, masonry-related structures have received less attention. This research seeks to implement an automated crack segmentation and a real-life crack length measurement of masonry walls using CV techniques and deep learning. The main contributions include (1) a large dataset of manually labelled images about various types of Korea masonry walls;(2) a careful performance evaluation of various deep learning-based crack segmentation models, including U-Net, deepLabV3+, and FPN;and (3) a novel algorithm to extract real-life crack length measurement by detecting the brick units. The experimental results showed that deep learning-based masonry crack segmentation performed significantly better than previous approaches and could provide a real-life crack measurement. Therefore, it has a huge po-tential for motivating masonry-based structure investigation.

关键词： Masonry building Crack segmentation deep learning Measurement image processing

来源：评论

学校读者我要写书评

暂无评论

Research on Obstacle Detection Algorithm of High Voltage Line Based on Improved YOLOv5 Algorithm

Research on Obstacle Detection Algorithm of High Voltage Lin...

引用

IEEE International Conference on Power, Intelligent Computing and Systems (ICPICS)

作者： Heng Zhou Qing Ai Jinghui Zhang Ruiting Zhang School of Intelligent Science and Engineering Hubei Minzu University Enshi Hubei China The State Grid Hubei Xuanen Enshi Hubei China

ISBN: (数字)9798350374315

ISBN: (纸本)9798350374322

Ensuring the secure and steady operation of high-voltage transmission lines is of paramount importance within contemporary power system operations and maintenance, given its direct impact on maintaining the continuity and dependability of electricity supply. real-time monitoring and identification of high voltage line obstacles is an effective means to prevent power failure and reduce economic losses. The meteoric advancement of deep learning technologies within the realms of image processing and object recognition has recently furnished a novel solution for identifying obstructions in high-voltage power lines. This study presents an enhanced You Only Look Once version 5 (YOLOv5)-based method tailored for detecting impediments along high-voltage transmission infrastructure. This method aims to improve the accuracy and efficiency of obstacle detection by optimizing the model structure and adjusting the training strategy. Firstly, we augment the fundamental YOLOv5 architecture by incorporating an attention mechanism, thereby amplifying its capacity to discern distinct features of obstacles. Subsequently, during the training phase, we deploy a learning rate modulation tactic to bolster the model's adaptability across diverse environmental contexts where high-voltage line obstructions may manifest. In addition, aiming at the diversity and complexity of obstacles in high-voltage lines, this study constructs a dataset containing large-scale data to ensure the comprehensiveness of model training. The experimental results show that compared with the traditional obstacle detection method and the unimproved YOLOv5 model, the proposed method in this study presents higher accuracy and faster processing speed in the obstacle detection task of high-voltage lines. This achievement not only demonstrates the application potential of deep learning in the field of obstacle detection in power systems, but also provides technical support for future high-voltage line monitoring and maint

关键词： YOLO Training deep learning Adaptation models Accuracy Attention mechanisms High-voltage techniques Maintenance Object recognition Context modeling

来源：评论

学校读者我要写书评

暂无评论

Weather Prediction Based on the Cloud images Using Machine and deep learning Algorithms

Weather Prediction Based on the Cloud Images Using Machine a...

引用

Next Generation Communication & Information processing (INCIP), International Conference on

作者： Rohil P. Utture Alan K. Biju Y.V. Srinivasa Murthy Linga Reddy Cenkeramaddi Nancy Kumari Department of Information Technology Manipal Institute of Technology Bengaluru Manipal Academy of Higher Education Manipal Karnataka India Department of Information and Communication Technology The ACPS Research Group The University of Agder Grimstad Norway School of Computer Science and Engineering VIT Bhopal University Madhya Pradesh India

ISBN: (数字)9798331528140

ISBN: (纸本)9798331528157

The process of weather forecasting based on the climatic images captured at various geographical locations and times can guide humans to plan their daily routines. In this paper, we propose forecasting the weather based on the images captured from the sky. The image dataset collected from various online search engines and mobile cameras pertains to timings such as sunrise, sunset, daytime, and night. Care has been taken to consider various weather conditions such as clear, foggy, rainy, and glare. Machine learning (ML) is widely adapted for various real-time applications. The inception of deep learning (DL) architectures has already been proven to outperform the traditional feature engineering approaches. This paper proposes two approaches. One is using the traditional ML models and the other is using trending convolutional neural networks (CNNs). In ML-based approach, decision tree (DT) and random forest (RF) are considered. Manually extracted features such as brightness, sharpness, contrast, white pixel intensity, haze factor, and color histogram are fed to them. Meanwhile, in the CNN-based approach, VGG16 and RESNET architectures are used. The performance of CNNs is found to be better when compared to the traditional ML algorithms. The basic CNN, VGG16, and RESNET models provide an accuracy of 62%, 72%, and 73.35%, respectively.

关键词： Support vector machines Radio frequency Accuracy Weather forecasting Predictive models Prediction algorithms Classification algorithms Convolutional neural networks Decision trees Random forests

来源：评论

学校读者我要写书评

暂无评论

Analysis of real-time Hostile Activitiy Detection from Spatiotemporal Features Using time Distributed deep CNNs, RNNs and Attention-Based Mechanisms

Analysis of Real-Time Hostile Activitiy Detection from Spati...

引用

International image processing, Applications and Systems Conference (IPAS)

作者： Labib Ahmed Siddique Rabita Junhai Tanzim Reza Salman Sayeed Khan Tanvir Rahman Department of Computer Science and Engineering BRAC University Dhaka Bangladesh Department of Computer and Information Sciences College of Engineering University of Delaware

ISBN: (纸本)9781665462204

real-time video surveillance, through CCTV camera systems has become essential for ensuring public safety which is a priority today. Although CCTV cameras help a lot in increasing security, these systems require constant human interaction and monitoring. To eradicate this issue, intelligent surveillance systems can be built using deep learning video classification techniques that can help us automate surveillance systems to detect violence as it happens. In this research, we explore deep learning video classification techniques to detect violence as they are happening. Traditional image classification techniques fall short when it comes to classifying videos as they attempt to classify each frame separately for which the predictions start to flicker. Therefore, many researchers are coming up with video classification techniques that consider spatiotemporal features while classifying. However, deploying these deep learning models with methods such as skeleton points obtained through pose estimation and optical flow obtained through depth sensors, are not always practical in an IoT environment. Although these techniques ensure a higher accuracy score, they are computationally heavier. Keeping these constraints in mind, we experimented with various video classification and action recognition techniques such as ConvLSTM, LRCN (with both custom CNN layers and VGG-16 as feature extractor) CNNTransformer and C3D. We achieved a test accuracy of 80% on ConvLSTM, 83.33% on CNN-BiLSTM, 70% on VGG16-BiLstm, 76.76% on CNN-Transformer and 80% on C3D.

关键词： deep learning Streaming media Feature extraction Cameras Video surveillance real-time systems Skeleton

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：